HMMG Covid19.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Mô hình Hidden Markov Gaussian\n",
    "\n",
    "\n",
    "### BS. Lê Ngọc Khả Nhi\n",
    "\n",
    "\n",
    "# Nhập đề\n",
    "\n",
    "Trong thời gian nằm viện, một vài thông số sinh lý của bệnh nhân thường được khảo sát lặp lại nhiều lần để theo dõi diễn tiến của bệnh lý, và cập nhật thông tin cho phép bác sĩ thay đổi chế độ điều trị phù hợp, kịp thời. Quy trình theo dõi này sinh ra dữ liệu dạng chuỗi (sequential data) chứa nhiều thông tin thú vị; tuy nhiên dạng dữ liệu chuỗi hiếm khi được khai thác trong nghiên cứu khoa học; hoặc nếu có thì mỗi thời điểm thường được khảo sát như những quan sát độc lập bằng repeated ANOVA, Mixed model, paired-samples t-test, những cách làm này đã bỏ qua yếu tố \"thời gian\", \"chuyển tiếp\" của dữ liệu chuỗi.\n",
    "\n",
    "Một trong những phương pháp thống kê thích hợp cho dữ liệu chuỗi đó là mô hình Markov ẩn (Hidden Markov model, viết tắt: HMM). Mô hình HMM giả định rằng giá trị quan sát là kết quả của một tiến trình Markov mang tính ngẫu nhiên (stochastic process) gồm các tham số trạng thái chưa xác định (trạng thái ẩn, hidden state X) với xác suất ban đầu (initial probability). Sự chuyển tiếp giữa các tham số trạng thái ẩn được xác định bằng xác suất chuyển tiếp (transition probability). Xác suất đầu ra (emission probability) cho phép liên kết mỗi trạng thái ẩn với kết quả quan sát được.\n",
    "\n",
    "Trong thí nghiệm hôm nay, Nhi sẽ áp dụng mô hình HMM Gaussian emission (HMMGE) với đầu ra là biến định lượng, giả định theo phân bố Gaussian) để khảo sát một bộ dữ liệu theo dõi độ bão hòa Oxy máu (SpO2) trong thời gian nằm viện của 35 bệnh nhân nhiễm COVID-19. Có thể xem đây là 1 bài toán Unsupervised learning.\n",
    "\n",
    "Chúng ta bắt đầu thí nghiệm nhé ?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Công cụ\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "import tensorflow as tf\n",
    "import tensorflow_probability as tfp\n",
    "\n",
    "# Đồ họa\n",
    "import matplotlib \n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from matplotlib import cm\n",
    "\n",
    "# network\n",
    "import networkx as nx\n",
    "from graphviz import Source\n",
    "\n",
    "# HMMG\n",
    "from hmmlearn.hmm import GaussianHMM"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Chuẩn bị dữ liệu\n",
    "\n",
    "Trước hết, Nhi tải dữ liệu gốc:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Patient</th>\n",
       "      <th>Time</th>\n",
       "      <th>Oxygen</th>\n",
       "      <th>Flow</th>\n",
       "      <th>SpO2</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>CC220447</td>\n",
       "      <td>09/04/2020 08:00</td>\n",
       "      <td>Non</td>\n",
       "      <td>0.0</td>\n",
       "      <td>93</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>CC220447</td>\n",
       "      <td>09/04/2020 05:00</td>\n",
       "      <td>Non</td>\n",
       "      <td>0.0</td>\n",
       "      <td>95</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>CC220447</td>\n",
       "      <td>08/04/2020 22:00</td>\n",
       "      <td>Non</td>\n",
       "      <td>0.0</td>\n",
       "      <td>97</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>CC220447</td>\n",
       "      <td>08/04/2020 08:00</td>\n",
       "      <td>Non</td>\n",
       "      <td>0.0</td>\n",
       "      <td>94</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>CC220447</td>\n",
       "      <td>08/04/2020 05:00</td>\n",
       "      <td>Non</td>\n",
       "      <td>0.0</td>\n",
       "      <td>95</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1204</th>\n",
       "      <td>RR000</td>\n",
       "      <td>27/03/2020 22:00</td>\n",
       "      <td>Oui</td>\n",
       "      <td>2.0</td>\n",
       "      <td>97</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1205</th>\n",
       "      <td>RR000</td>\n",
       "      <td>27/03/2020 20:00</td>\n",
       "      <td>Oui</td>\n",
       "      <td>2.0</td>\n",
       "      <td>97</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1206</th>\n",
       "      <td>RR000</td>\n",
       "      <td>27/03/2020 16:00</td>\n",
       "      <td>Oui</td>\n",
       "      <td>3.0</td>\n",
       "      <td>97</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1207</th>\n",
       "      <td>RR000</td>\n",
       "      <td>27/03/2020 12:00</td>\n",
       "      <td>Oui</td>\n",
       "      <td>2.0</td>\n",
       "      <td>98</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1208</th>\n",
       "      <td>RR000</td>\n",
       "      <td>27/03/2020 08:00</td>\n",
       "      <td>Oui</td>\n",
       "      <td>3.0</td>\n",
       "      <td>98</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1209 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       Patient              Time Oxygen  Flow  SpO2\n",
       "0     CC220447  09/04/2020 08:00   Non    0.0    93\n",
       "1     CC220447  09/04/2020 05:00   Non    0.0    95\n",
       "2     CC220447  08/04/2020 22:00   Non    0.0    97\n",
       "3     CC220447  08/04/2020 08:00   Non    0.0    94\n",
       "4     CC220447  08/04/2020 05:00   Non    0.0    95\n",
       "...        ...               ...    ...   ...   ...\n",
       "1204     RR000  27/03/2020 22:00   Oui    2.0    97\n",
       "1205     RR000  27/03/2020 20:00   Oui    2.0    97\n",
       "1206     RR000  27/03/2020 16:00   Oui    3.0    97\n",
       "1207     RR000  27/03/2020 12:00   Oui    2.0    98\n",
       "1208     RR000  27/03/2020 08:00   Oui    3.0    98\n",
       "\n",
       "[1209 rows x 5 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "path = 'covidspo2.csv'\n",
    "df = pd.read_csv(path, sep =';')\n",
    "\n",
    "df"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Quan sát cho thấy dữ liệu có dạng bảng dọc (Long table format), trong đó mỗi bệnh nhân được mã hóa bằng 1 chuỗi string (cột Patient), và liên kết với 3 chuỗi: Oxygen = biến định tính, Có thở oxy hay không, Flow = Lưu lượng oxy cung cấp; và SpO2. Nhi cũng nhận ra là trình tự của chuỗi bị đảo ngược (từ ngày cuối ngược dần về ngày đầu nhập viện) - xem cột Time.\n",
    "\n",
    "Để đơn giản, ta chỉ quan tâm đến 1 chuỗi là SpO2; đơn vị quan sát của mô hình HMM là chuỗi nên Nhi cần tạo ra 1 database dạng dictionary, trong đó keys = tên mỗi bệnh nhân là duy nhất, tương ứng với value = 1 array ghi giá trị SpO2 theo trình tự thời gian xuôi (nhập viện-xuát viện). "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_seq_db(path = 'covidspo2.csv', target = 'SpO2'):\n",
    "    df = pd.read_csv(path, sep =';')\n",
    "    df['Time'] = pd.to_datetime(df['Time'])\n",
    "\n",
    "    patients = np.unique(df['Patient'])\n",
    "    seq_db = dict.fromkeys(patients)\n",
    "    \n",
    "    for p in patients:\n",
    "        seq_db[p] = np.array(df[df['Patient'] == p][target].iloc[::-1])\n",
    "        \n",
    "    return seq_db"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Áp dụng hàm create_seq_db, Nhi có database cần thiết:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{'BT000000': array([96, 98, 97, 98, 98, 98, 97, 96, 98, 97, 96, 95, 98, 94],\n",
       "       dtype=int64),\n",
       " 'CBCor': array([96, 92, 89, 85, 95, 88, 93, 94, 94, 94, 88, 87, 92, 84, 94, 91, 87,\n",
       "        64, 67], dtype=int64),\n",
       " 'CC220447': array([91, 89, 73, 94, 93, 95, 94, 97, 95, 93], dtype=int64),\n",
       " 'DBV000000': array([97, 94, 94, 94, 94, 96, 94, 94, 95, 97, 97, 96, 94, 96, 95, 96],\n",
       "       dtype=int64),\n",
       " 'DE230768': array([94, 94, 96, 94, 95, 96, 93, 94, 95, 95, 94, 94, 90, 94, 95, 97, 97,\n",
       "        98, 99, 97, 99, 99, 97, 98, 97, 99, 98, 98, 99], dtype=int64),\n",
       " 'DJ000000': array([97, 93, 94, 95, 98, 97, 99, 99, 92, 94, 92, 95, 96, 96, 93, 94, 94,\n",
       "        91, 97, 97, 90, 93, 91, 99, 97, 96, 90, 91, 95, 85, 88, 91, 90, 80,\n",
       "        90, 83, 70, 83, 90, 82, 90, 77, 73, 74, 96, 74, 73, 69],\n",
       "       dtype=int64),\n",
       " 'FC000000': array([96, 97, 97, 94, 97, 95, 97, 96, 97, 99, 96, 96, 97, 96, 97, 98, 95,\n",
       "        96], dtype=int64),\n",
       " 'FJ000000': array([95, 93, 94, 94, 95, 97, 93, 97, 97, 98, 97, 95, 97, 98, 96, 95, 91,\n",
       "        97, 97, 96, 96, 97], dtype=int64),\n",
       " 'FJQ160143': array([98, 98, 96, 91, 96, 92, 94, 96, 96, 94, 96, 95, 95, 96, 93, 95, 94,\n",
       "        90, 96, 93, 92, 93, 95, 92, 90, 95, 95, 96, 93, 97, 96, 97, 97, 97,\n",
       "        94, 95, 96, 97, 94, 94, 96, 97, 96, 96, 97, 97, 96, 97, 95, 93, 96,\n",
       "        97, 92, 93, 92], dtype=int64),\n",
       " 'FN230635': array([96, 95, 95, 95, 97, 95, 95, 94, 94, 95, 94, 96, 96, 94, 93, 92, 93,\n",
       "        93, 95, 93, 94, 90, 93, 98, 97, 97, 98, 94, 96, 91, 96, 93, 91, 92,\n",
       "        91, 97, 98, 93, 92, 90, 92, 97, 95, 92, 93, 91, 88, 96, 96, 96, 91,\n",
       "        94, 94, 95, 93, 97, 90, 91, 89, 94, 96, 92, 97, 93, 96, 97, 97, 96,\n",
       "        96, 94, 93, 97, 94, 95, 94, 94, 95, 97, 97, 95], dtype=int64),\n",
       " 'GC000000': array([92, 90, 93, 97, 95, 96, 96, 96, 89, 97, 95, 96, 98, 97, 94, 92, 95,\n",
       "        96, 98, 94, 96, 94, 95, 97, 91, 91, 95, 93, 96, 96, 93, 90, 88, 91,\n",
       "        91, 90, 89, 95, 87, 88], dtype=int64),\n",
       " 'GF000000': array([ 97,  94,  97,  95,  92,  92,  92,  94,  95,  93,  95,  95,  97,\n",
       "         94,  96,  94,  95,  97,  97,  94,  93, 100,  96,  97,  93,  93,\n",
       "         97,  95,  94,  95], dtype=int64),\n",
       " 'GMA000': array([92, 91, 91, 95, 90, 89, 95, 91, 97, 94, 91, 93, 94, 96, 96, 95, 95,\n",
       "        91, 88, 95, 94, 95, 93, 96, 95, 94, 98, 91, 90, 94, 90, 93, 95, 94,\n",
       "        93, 93, 93, 93, 92, 93, 95, 95, 93, 94, 94, 95, 93, 94, 94, 96, 76,\n",
       "        93, 98, 98, 96, 95, 98, 98, 95, 96, 95, 97, 95, 95, 97, 96, 94, 96,\n",
       "        97, 95, 96, 95, 96, 93, 97, 97, 96, 98, 93, 93, 94, 93, 94, 92, 95,\n",
       "        93, 88, 83, 90, 88, 79, 76, 79], dtype=int64),\n",
       " 'GT000000': array([96, 98, 96, 95, 95, 93, 96, 92, 95, 97, 97, 98, 95, 94, 98, 95, 95,\n",
       "        96, 95, 97, 97, 97, 94], dtype=int64),\n",
       " 'HA000': array([91, 93, 89, 87, 87, 93, 89, 89, 88, 90, 90, 86, 89, 92, 88, 94, 91,\n",
       "        96, 86, 98, 98, 95, 69, 94, 98, 93, 93, 99, 97, 99, 93, 94, 90, 93,\n",
       "        95, 96, 98, 95, 98, 93, 94, 89, 93], dtype=int64),\n",
       " 'HJ100230': array([90, 97, 94, 98, 96, 96, 94, 99, 99, 92, 92, 97, 96, 97, 98, 98, 93,\n",
       "        96, 95], dtype=int64),\n",
       " 'JR281044': array([91, 97, 92, 92, 97, 95, 91, 95, 95, 94, 89, 98, 95, 95, 95, 92, 93,\n",
       "        90, 95], dtype=int64),\n",
       " 'LJQ000000': array([84, 89, 89, 89, 88, 90, 87, 90, 90, 85, 90, 88, 90, 90, 92, 90, 90,\n",
       "        92, 91, 92, 92, 92, 94, 94, 93, 93, 93, 92, 92, 93, 93, 90, 91, 94,\n",
       "        90, 91, 94, 84, 90, 89, 91, 92], dtype=int64),\n",
       " 'LL150837': array([91, 92, 90, 90, 92, 96, 98, 96, 96, 96, 97, 94, 93, 94, 95, 93, 95,\n",
       "        91, 93, 96, 93, 91, 94, 94, 91, 93, 94, 97, 96, 92, 91, 92, 93, 90,\n",
       "        92, 88, 94, 80, 93, 94, 93, 93, 92, 91, 96, 94, 84, 94, 94, 91, 94,\n",
       "        93, 95, 93, 94, 97, 97, 96, 95, 91, 95, 93, 93, 93, 95, 94, 96, 96,\n",
       "        93, 93, 93, 96, 94, 94, 92, 92, 92], dtype=int64),\n",
       " 'LL171069': array([92, 92, 95, 94, 93, 90, 91, 92, 90, 93, 94, 92, 94, 93, 92, 98, 95,\n",
       "        94, 93, 93, 94, 96, 98, 92, 95, 95, 95, 95, 96, 94, 94, 95, 97, 94,\n",
       "        94, 92, 94, 94, 96, 93, 94, 95, 94, 93, 94, 94, 94, 93, 93, 94, 91],\n",
       "       dtype=int64),\n",
       " 'MC0000': array([89, 94, 94, 96, 95, 95, 97, 92, 98, 95, 95, 95, 96, 94, 97, 91, 97,\n",
       "        95], dtype=int64),\n",
       " 'ME000': array([94, 96, 89, 89, 90, 93, 91, 93, 92, 91, 92, 91, 93, 92, 93, 92, 93,\n",
       "        93, 92, 93, 94, 94, 94, 95, 93, 95, 96, 94, 94, 93, 95, 95],\n",
       "       dtype=int64),\n",
       " 'MJJ000000': array([89, 89, 89, 96, 84, 89, 85, 83, 85, 94, 85, 77, 89, 86, 76, 78, 87,\n",
       "        66, 61], dtype=int64),\n",
       " 'MN000': array([94, 90, 91, 94, 93, 93, 95, 96, 63, 93, 90, 92, 94, 95, 94, 94, 93,\n",
       "        95], dtype=int64),\n",
       " 'MV000000': array([97, 97, 99, 94, 92, 96, 95, 96, 95, 96, 97, 94, 95, 96],\n",
       "       dtype=int64),\n",
       " 'NM000000': array([85, 81, 95, 97, 95, 96, 87, 90, 91, 81, 81, 90, 90], dtype=int64),\n",
       " 'PG110767': array([93, 98, 97, 94, 95, 96, 90, 95, 96, 96, 91, 93, 94, 96, 94],\n",
       "       dtype=int64),\n",
       " 'PMA000': array([96, 98, 97, 96, 99, 97, 97, 96], dtype=int64),\n",
       " 'PY000000': array([93, 97, 96, 94, 95, 98, 97, 97, 97, 97, 95, 96, 95, 95, 95, 95, 96,\n",
       "        96, 96, 96, 95, 98, 95, 98, 97, 96, 95, 93, 96, 96, 93, 91, 94, 99,\n",
       "        96, 96, 94, 93, 94, 95, 91, 97, 93, 97, 91, 94, 92, 91, 95, 93, 91,\n",
       "        84, 89, 90, 99, 78, 82, 80, 75], dtype=int64),\n",
       " 'RF100266': array([99, 97, 98, 99, 99, 99, 99, 99, 99, 98, 98, 99, 97, 98, 98, 95, 97,\n",
       "        96, 97, 97, 98, 97, 97, 96, 95, 98, 97, 96, 97, 98, 97, 97, 96, 97,\n",
       "        96, 97, 98, 96, 96, 97, 96, 96, 96, 95, 96, 94, 97, 95, 97, 97, 96,\n",
       "        97, 96, 96], dtype=int64),\n",
       " 'RR000': array([98, 98, 97, 97, 97, 98, 96, 94, 98, 98, 98, 98, 98, 98, 98, 98, 96,\n",
       "        98, 98, 98, 99, 95, 96, 95, 97, 97], dtype=int64),\n",
       " 'SJ000000': array([98, 98, 94, 95, 98, 98, 97, 97, 99, 98, 97, 99, 97, 97, 98, 99, 97],\n",
       "       dtype=int64),\n",
       " 'SJ110834': array([89, 95, 90, 90, 91, 92, 93, 90, 90, 86, 92, 94, 90, 94, 94, 93, 91,\n",
       "        95, 90, 90, 90, 95, 90, 90, 93, 92, 93, 92, 92, 91, 91, 94, 96, 91,\n",
       "        91, 92, 99, 86, 91, 77, 90, 91, 89, 87, 90, 90, 92, 94, 90, 90, 89,\n",
       "        89, 94, 94, 90, 91, 89, 90, 91, 91, 90, 95, 95, 94, 93, 92, 91, 94,\n",
       "        98, 92, 91, 94, 96, 94, 94, 95, 95, 95, 93, 91, 93, 89, 96, 94, 92,\n",
       "        93, 96], dtype=int64),\n",
       " 'VDB000': array([97, 94, 94, 94, 94, 96, 94, 94, 95, 97, 97, 96, 94, 96, 95, 96],\n",
       "       dtype=int64),\n",
       " 'WJ151131': array([88, 96, 96, 94, 91, 93, 94, 95, 94, 91, 94, 97, 95, 93, 97, 97, 95,\n",
       "        95, 97, 96, 93, 97, 97, 95, 99, 98, 93, 94, 92, 91, 95, 98, 97, 98,\n",
       "        98, 90, 96, 98, 98, 95, 98, 97, 97, 96, 95, 96, 97, 98, 95, 94, 96,\n",
       "        94, 97, 97, 97, 92, 91, 92, 94, 96, 95, 93, 93, 93, 94],\n",
       "       dtype=int64)}"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "seq_db = create_seq_db('covidspo2.csv')\n",
    "\n",
    "seq_db"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Danh sách bệnh nhân và độ dài chuỗi SpO2 của họ:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Patient</th>\n",
       "      <th>Len</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>BT000000</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>CBCor</td>\n",
       "      <td>19</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>CC220447</td>\n",
       "      <td>10</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>DBV000000</td>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>DE230768</td>\n",
       "      <td>29</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>DJ000000</td>\n",
       "      <td>48</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>FC000000</td>\n",
       "      <td>18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>FJ000000</td>\n",
       "      <td>22</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>FJQ160143</td>\n",
       "      <td>55</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>FN230635</td>\n",
       "      <td>80</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>GC000000</td>\n",
       "      <td>40</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>GF000000</td>\n",
       "      <td>30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>GMA000</td>\n",
       "      <td>93</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>GT000000</td>\n",
       "      <td>23</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>HA000</td>\n",
       "      <td>43</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>HJ100230</td>\n",
       "      <td>19</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>JR281044</td>\n",
       "      <td>19</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>LJQ000000</td>\n",
       "      <td>42</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>LL150837</td>\n",
       "      <td>77</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>LL171069</td>\n",
       "      <td>51</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>MC0000</td>\n",
       "      <td>18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>ME000</td>\n",
       "      <td>32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>MJJ000000</td>\n",
       "      <td>19</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>MN000</td>\n",
       "      <td>18</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>MV000000</td>\n",
       "      <td>14</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>NM000000</td>\n",
       "      <td>13</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>PG110767</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>PMA000</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>PY000000</td>\n",
       "      <td>59</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>RF100266</td>\n",
       "      <td>54</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>RR000</td>\n",
       "      <td>26</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>SJ000000</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>SJ110834</td>\n",
       "      <td>87</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>VDB000</td>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>WJ151131</td>\n",
       "      <td>65</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      Patient  Len\n",
       "0    BT000000   14\n",
       "1       CBCor   19\n",
       "2    CC220447   10\n",
       "3   DBV000000   16\n",
       "4    DE230768   29\n",
       "5    DJ000000   48\n",
       "6    FC000000   18\n",
       "7    FJ000000   22\n",
       "8   FJQ160143   55\n",
       "9    FN230635   80\n",
       "10   GC000000   40\n",
       "11   GF000000   30\n",
       "12     GMA000   93\n",
       "13   GT000000   23\n",
       "14      HA000   43\n",
       "15   HJ100230   19\n",
       "16   JR281044   19\n",
       "17  LJQ000000   42\n",
       "18   LL150837   77\n",
       "19   LL171069   51\n",
       "20     MC0000   18\n",
       "21      ME000   32\n",
       "22  MJJ000000   19\n",
       "23      MN000   18\n",
       "24   MV000000   14\n",
       "25   NM000000   13\n",
       "26   PG110767   15\n",
       "27     PMA000    8\n",
       "28   PY000000   59\n",
       "29   RF100266   54\n",
       "30      RR000   26\n",
       "31   SJ000000   17\n",
       "32   SJ110834   87\n",
       "33     VDB000   16\n",
       "34   WJ151131   65"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "lengths = [len(seq_db[p]) for p in seq_db.keys()]\n",
    "\n",
    "pd.DataFrame({'Patient':list(seq_db.keys()), 'Len': lengths})"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Ta quan sát một vài chuỗi SpO2 của các bệnh nhân chọn ngẫu nhiên:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "def plot_seq(patient = 'GMA000', color = 'red'):\n",
    "    spo2 = seq_db[patient]\n",
    "    matplotlib.rcParams.update({'font.size': 15})\n",
    "    plt.figure(figsize=(24,5))\n",
    "    \n",
    "    plt.title(f'SpO2 sequence of patient {patient}')\n",
    "    plt.fill_between(np.arange(spo2.shape[0]),\n",
    "                     np.repeat(np.min(spo2)-2, spo2.shape[0]),\n",
    "                     spo2,\n",
    "                     color = color,\n",
    "                     step=\"pre\", \n",
    "                     alpha=0.3)\n",
    "    plt.xlabel('Time points')\n",
    "    plt.ylabel('SpO2')\n",
    "\n",
    "    plt.plot(np.arange(spo2.shape[0]),\n",
    "             spo2,\n",
    "             color = color,\n",
    "             drawstyle=\"steps\", \n",
    "             alpha = 0.4)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAABXYAAAFbCAYAAABxtdYPAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjEsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+j8jraAAAgAElEQVR4nOzde5glZ1kv7N9DAkKQgZAEGHRC1AEyGhQhKAdFzkgQEPAQFDHI4VNAVHCDIITDFmFA5AwxbmNETio7GAYngkGQsEMYEgygDJFRAiQDgRlDQjpAMLzfH1VNFovVM91rpld3dd/3da1rpaveqnpWVXWl5tfveqtaawEAAAAAYDiut9IFAAAAAACwNIJdAAAAAICBEewCAAAAAAyMYBcAAAAAYGAEuwAAAAAAAyPYBQAAAAAYGMEuAAAcgKq6d1WdV1VfrapWVSetdE3jquriqnr/StcBAMDBI9gFANaMqvrBqjq1qj5VVVdX1eVV9cmq+ququvdBWP8tquqlVfXvVTXXB3n/WlUnV9WGCe1vV1Uv7EO/L/ftL6yqP6yqGx9oPay8qjo8yRlJbpzk6Ul+LckHVqiW51fVz6/EtsfquGNfyzFTLHurqnpRVV1QVV+pqm9W1Zeq6r1V9ftVdcRY+9P7ML1V1fELrPP3RtqctECbQ6pqd9/mOfup8fZV9ff99WWuqs6pqvss0PZ6/fY/VVVfr6rPV9XLF/r9X8q6AQCqtbbSNQAAHLA+1PmXJN9M8sYk/57kRklul+QhSc5srT3lANZ/tyTbkmxI8uYk5yU5JMm9kzwyyWeTPLC19h8jy7wkyZOTvLNv/82+/S8l+XiSu7bWvjZtTay8qnpAkncneWRr7YwVrqUl+avW2kkT5n1PktZau2YGdZyU5C+T3Lu19v4lLPezSd6W5LB0Yfl5Sa5IcmSSuyV5cJKLW2u3H1nm9CS/nuTrSU5rrT15wno/nuS2SW6Y5LGttdMntPm5dL/f/5mkkmxuE/6hVFU/lGRHkv9J8sq+vickOS7Jg1prZ4+1f1WSpyZ5R5KzkmxJ8ttJzklyv9bat6ZdNwDAoStdAADAQfK8dIHQj7fWLhydUVVPSXKraVdcVbdKcma6e6d7tNY+MjL79VV1QpK/T/LOqvrxkbD27Ule3Fq7YqT9KVX16SR/mORxSV47bV2sCvPn1X+vaBX70Vr7xkrXsC9V9SNJ/m+SvUnu1lrbOaHNLdOFpJO8I8mjquppo5+1qu6S5A5J3pLkV/ZRwuOS/FeSp6X7Xb9XkvdNaPfiJDdLcuf560xVzf8h6XVVdex8INx/pt9OckZr7ZEjNX0myauTnNjXteR1AwAkhmIAANaO2ybZOx7qJklr7Vuttd2j0/qvXJ9eVffrh0q4uqq+WFWvmvA16f+V5KgkzxoLdefXvz1dD7vbpwuI5qefPxbqzvub/v24xXywqnpMVe3ov5o+V1X/VVVvrqqjxtrdtqr+uqq+UFXX9OOqvmzS176r6qeq6v9V1deq6rKqem1VHdfvl+ePtDupn3avCet4f1VdPGH68VX1jqraU1XfqKqL+uEnDp20fFXduqreOvL183dX1e0mrPcGVfWM6oazuLqqrqiq8/vgfrTdTatqa1Xt6rf/5X79P7iI3T2/jh/tP8Pe/iv0n+y3fchIm4uT/FX/4/vmv+6/n/XODx1wVFW9sV//XHVDDfz4hPZPqqr3VNWl/TH9QlW9qUaGOaiqY0a2++t13bADbaTNxDF2D/ax6s+dvxzfJ9X1rN2XF6b7w8zjJoW6SdJau6y19ocLLH9aksOTjA9F8dgkX07yroU2XFW3SNcb+K+SbE/ypYz8Ho+0u3GShyZ5/+h1prV2VZL/k+7bAXcZWeRR6Xr/vnJsVX+e5Ookjz6AdQMA6LELAKwZ/5nk9lX1iCV8Jf5OSX4hXdDyxnTDJDw1yXFVdf+Rr0k/Msk1uS7Em+TP0wXAj8z+e+F+f/9+2f4KrKpH99s9J8nJSb6W5OgkD0pyi3ShVarqzkn+OclXkvxZkkuT/Fj/ee5RVT/TWvtm3/Ynk5yd5KtJtvbLnNjvgwNSXe/ldyTZleTl6Xqy3i1dcHfHJL84tsiN041Je16SZyf5gSS/k+TMqjqutXZtv94bpBvy4F5J3pPkTem+fn+HJI9Iv8+r6qZJzu330WnpejtuTPKkJB+uquNba5/dz2cYHdbjdUm+mG44j63p9umv9k1/N91xeGKSP04yMZBcwD/2++b56Xr9PiXJB6rqbq21fxtp9/v9vnl13/64JI9Pcp+qukNrbW+6c+DXkvx1uvPk1MUUsEzH6ox0+3t8n/znPuq4Ybpg9bOttX9aTO0TXNi/fiP9H0769T4qXdD8zX0s++vp/l301621/6mqNyf5zaq66dgfZn40yfck+dCEdZzXv98l3XAK8//9rZGfkyStta9X1YX5zqB2qesGABDsAgBrxh8luX+S/1vdUAcfTPKRdD3gFgrc7pDk4a21v+9/fn1dNybmLyV5W1XdJMltknyitXb1QhtvrX26qr7ar3NBfY/Pk9ONo/mWfbXtPSJdAHuf1tr/jEx/7li705J8IcldWmtfHdnee9OFbb+a5PR+8ivSfXPrHvNjAlfV69Pts6n1QdppST48Vu+fVdXHkvxpVd1rbNzVI5O8rLX20pH1fDnJS5PcL12Ym3Qh6r3SDW3x7LHtjn4L7YVJfjDd+MUfG2lzepJPJHlBkpP281FelS5ku1tr7eP98q9NFxj+SlWd1lp7b2vt76vqZulCzH9ayniy6cZkfuTI1/bPSHe+/kmSnx1pd4fW2tzY531numD+cUle2s9/U1X9dZL/aq29aX8bX65j1Vr7eFV9KEvbJ7dNt78/Nj6jr/N7xyZ/Zex3Yd5pSV5ZVZtaa59P97tzs376sfvY/mOTfKC19pn+59OT/F66oRveMNLu1v37pRPWMT/t+8ba71lgGIxLk9y9qm7Qj3u81HUDABiKAQBYG1prH0py53S9W2+aLqx5fZJPVvdk+Ulfw79oJNSd95L+/eH9+4b+fdKQCuOu7Le9L69MctckJ7fWLlrEOq9I9xX1B1dVTWpQVXdI1+PvLUm+p6qOnH+lC2vnkjygb3uLdL0yzxx90FsfLr1iEfXsy/2T3DJdD8mbjdWxvW/zgLFlvpWuN+qof+7fbzsy7VeTXJ4uuP0O8z2r+/3zq+l6lV46tv25dD0fx7f/Hfr9c/ck75wPdftttHQ9UJPrzo0D8dLR8VJbaxck+ack96uq7x2ZPtfXdb3qhpg4Ml0AekWSnzyA7S/nsVqq+d+xKyfMe3y6HsmjrzsusJ43p+uZ+5j+58cm+chYD+jvUFV3T/dAs2/3xu+P+3zv31GH9e+Tgtqvj7WZ/++FxjYeb7/UdQMA6LELAKwdrbVPpO+NWVW3SfIz6YKhn073dfE79wHmvO/qydta+0JVfSVdr8/kurBpf4Ft0gVUCwbAVfW/033l/tTW2osXsb6kCxPvme7hbHur6l+SnJXkb0Z65m7p31/Qvya5Zf8+/7k+NaHNJxdZ00Lm6zhtH21uOfbz7tba18em7e3fjxiZdtskF05oO+qofpkHpB+iYoJvLTB93g/07/8+Yd4n++UXPVbvPkzqRf7JdLXfZn77VXWfdD28fzLJDcfaH34A21/OY7VU879jGybM+/tcd64+Jt2QExO11v677818UlW9Kcl90v2+7cvj0oXB/1pVm0emvzvJM6vqR0cC/vke+98zYT03HGsz/9+3WGC74+2Xum4AAMEuALA29eOovrH/evo5Se6R5CfyncMNLPSgq2/3jG2tfbWqPpdu/N7DFhqOoQ+FbpLk/QvMf36S56TrIfmbS/gcn66qH05y3/71M+nG831BVd2ztfafI/W+PN3YrZNcPvbZJn32ST2C9/UwsPF7yfnl/1e6Ho+T7B77+dp9rH+8nn0+mGyk/dnpxsOdxsRe0TPyHduuqrukG094V5I/SPKZdGMstyRvy4F9+265j9VSfDpdT9Xv6onbWrskySVJ98C/RazrtHR/+PjzdONiv3Whhn3P6F9Kcv0k/7pAs99INwxIct3+mDQkwvy00aEUdif54ar6ngnDMXxfumEarhlpu5R1AwAIdgGAta211qrqw+mC3fHQ5IfH21fVxnS9c/9rZPIZ6cKdxyQ5ZYFNPX6k7fg6n5fkeekeTvb40a/gL/IzfCPd1+O39+s7Ick/JHlakienC8aS5NrW2tn7Wd38Q6y2TJg3adp/9+83nzDvB/KdD6War2NuEXUs1X8k2bJASDbvy+keBLfhALY/f9x/ZMK8Y9OFqf81Yd5Sbcl1D8UanXZtuvF3k26M10OSPGhk/NdU1Y1zYL11k+U9Vks9v79eVf+Q5BH9QwunfYBa0gXhl6QbauItrbWv7KPtL6Ubv/fZuW5/jHpqkkdX1TP6APYT6QLou01oe9f+/fyRaR9J1wP7J9L9cSnJt8cNvmO6IUPmLXXdAADG2AUA1oaqun9VfdcfravqRrlurNDxoQZuX1U/Pzbtmf376Ni7L0v3lfMXV9WdJmzjgelC1v9I8hdj805O8vwkf53ksfPjwS5WP+bpuI/27/Nh678m+bckvzlpLOGqOrSqbp4krbUvpQsUH1ZVtxtpc4N0D4waNz8O7/3G1vmoXPfAp3nvTvKlJH8wv72xZW7UP4xuGm9OF2Y+Z8J6K/n2WLtvTvITVfULk1bSj6G7oH7/nJvkIVV13Ng2ntX/+I5pPsCYZ4yOmdyfV/dL8t7W2lX95PkesuO9YZ+dyffxV2VyAD/Jch6r+foXW0vSDTdxdZK/qKpJf2BIFtEruD8HnpxuSJL99dp+XLo/XLystfb28Ve63+UjkjysX/dVSbYluVdV/di3i+p6/j4+XTi8Y2T9f5Mu5P7dfKcnpBsv980jdS913QAAeuwCAGvGK5Ic0Y+x+Yl0IdGmdL0eb5fkjf0YvKM+keRNVfXn6YKTeyf5hST/ki6USZK01nb3AfCZST5UVW9OF44ekuRe/TKfS/LQ0aEaqmo+YPpcuuEBfmXs+WeXLaJ34nuq6op0vfs+n+Rm6cYRbunC4vleyb+W7kFWH6+q09KN0XpYks1JHpEulDy9X+fT0g0Z8f+q6nXpermemAn3hq21i6rq7CT/Xx9EXpiut+HD0w0RcP2RtnNV9Zh0ofhFfR27+pqP7et4eBYYrmI/XpXkIUmeMzJEwdfT9ay9fa4Lnv8wXe/sv62qv013nK5JN27tCUkuSD8O8z78Trpz4Jx+/3wxyc8leWC6XqDvnaL+cbdJ8u7+fN2YbizYr6UbGmHeO9KF7dur6tT+c9w/3YPy9kxY53npHr72zHTnXGutvW3Sxpf5WH0k3VjEf1hVh6d7cN1nWmsfXmiB1tq/V9Uj0w0x8bGqOiPJh9KNv3tUkrukC1ivyHXDiiy0rncmeee+2lTVsekeknd6a+1/Fmj2znQ90h+X5O/6ac9KNyTKe6rqFX19T0j3bYAHjz0Q7xP9+fOU/vNsT9cr+6npzq+3jG1v0esGAEiStNa8vLy8vLy8vAb/Stcr93VJPpYu9PqfdL1s35dunMzrjbVv6YLO+yX5cLpQ7bIkr0lykwW2caskf5LuwVdXp+uZeGG6YRZuOqH96f12Fnq9fxGf6wlJ/ilduHhNki+kC4juPaHtbdINFXFx33ZvuiDzxUk2jbW9Z7qeqV9P13PzdUmO6+t6/oTP/Xfpgqar0o1huiVd6HfxhDqOS/KmdGOCXtPv13OTPDfJzUfaLbT8MQvUccN0we2/93V/JV2I+KSxdof12/pEf1y/2h+zP0/yk4s8n34sXej53+m+Ir8zyTOSHDLW7qS+1nstcr3z58RR6YL5vf259M9J7jyh/c/3x3Au3Xn9tiRH98f4/WNtb5su8L5y/hwbmfdd7Zf5WP16uh7y1/TzT1/k/tmY5EXpeqVfkS5Y/VK/f34/yREL7M8j97PeX+jbndT//LL+54fsZ7l3p+s5vWlk2pZ0f+T5Sn/sPpjkfgssf0iSpye5qD+PLk3yp0m+d4H2i163l5eXl5eXl1e15g+/AMD6U1UtyV+11k5a6VpWi6o6Jt0Dul7QWnv+ihazRlXV6Ul+vbW2kg9pAwBgDTDGLgAAAADAwAh2AQAAAAAGRrALAAAAADAwxtgFAAAAABgYPXYBAAAAAAbm0JUu4GA78sgj2zHHHLPSZQAAAAAAHJALLrhgT2vtqEnz1lywe8wxx+T8889f6TIAAAAAAA5IVX12oXmGYgAAAAAAGBjBLgAAAADAwAh2AQAAAAAGRrALAAAAADAwgl0AAAAAgIER7AIAAAAADIxgFwAAAABgYAS7AAAAAAADI9gFAAAAABiYQ1e6AABYbjt3Jrt2LX25zZuTLVsOfj0Aa8W019dpzOqavBY/E7Cy3IsCy0WwC8Cat2tXctZZSWuLX2ZuLtm4Mdm6dfnqAhi6aa6v05jlNXktfiZgZbkXBZaLYBeAdaG15IQTFt9+x45kz57lqwdgrVjq9XUas74mr8XPBKws96LAcjDGLgAAAADAwAh2AQAAAAAGRrALAAAAADAwgl0AAAAAgIER7AIAAAAADIxgFwAAAABgYAS7AAAAAAADI9gFAAAAABiYQ1e6ABianTuTXbuWvtzmzcmWLQe/HgAAONhW+z3vaq9vGmvxM83KtPtuWvY5sFoIdmGJdu1KzjoraW3xy8zNJRs3Jlu3Ll9dAABwsKz2e97VXt801uJnmpVp9t207HNgNRHswhRaS044YfHtd+xI9uxZvnoAAOBgW+33vKu9vmmsxc80K0vdd9Oyz4HVxBi7AAAAAAADI9gFAAAAABgYwS4AAAAAwMAIdgEAAAAABkawCwAAAAAwMIJdAAAAAICBEewCAAAAAAzMoStdADBMO3cmu3bNZlubNydbtsxmW0xn2vPBsQXYN9fX68zNJdu2LX25tbYvnBMHZprzaKn7btpjtHfv0peZxizPobV6vi71PJrVsQXWH8EuMJVdu5KzzkpaW97tzM0lGzcmW7cu73Y4MNOcD44twP65vnY2bkx27062b1/acmtxXzgnpjfNeTTNvjuQ++QNG5a+zFLN8hxai+frtNejWRxbYP0R7AJTay054YTl3caOHcmePcu7DQ6OpZ4Pji3A4ri+Jps2da+lWov7InFOTGua82jafTeL++QDMctzaK2dr9NejwCWgzF2AQAAAAAGRrALAAAAADAwMw92q+rnq+rjVfWNqvpMVT1tQpuLq6qNvb4461oBAAAAAFajmY6xW1X3SHJGktOS/H6Sn0yytaq+1Vp75VjztyR5zcjP18ymSgAAAACA1W3WD087OckHW2uP739+T1UdnuTkqnp9a200vP1Ca+28GdcHAAAAALDqzXoohjsmOXts2nuSHJ7kbjOuBQAAAABgkGYd7N4w3z2kwjf69y1j03+jqq6pqiuq6u1VdZvlLw8AAAAAYPWb9VAMu5LcZWzaT/TvNx+ZdmaS85Jcki7wfV6Sc6rqDq21K5a9SibauTPZtWvpy23enGwZj+0PstVcGwdubi7Ztm0221rN52uy9PpW++/GtMfW7+7szPJ8ZXoHcpymsVqPrfOVSZb6/5q9e5evlnHT/H9w2vpmtR9mdd82y+O02q3mY+s4sRLcD7CezDrYPSXJG6rqCUneni7UfXo/79r5Rq213xlZ5pyqOjfJhUkem2T8IWupqicmeWKSHH300ctTOdm1KznrrKS1xS8zN5ds3Jhs3bp8dSWruzYOzMaNye7dyfbty7+t1Xy+JtPVt5p/N6Y9tn53Z2uW5yvTm/Y4TWM1H1vnK+Om/X/Nhg3LU8+oA7nHWWp9s9oPs7xvS2ZznFa7IRxbx4lZcz/AejLrYPe0JD+W5A1JTk1ydZJnJnlNkssWWqi19m9VdVGSOy0w/9R+fTn++ONn8E+a9au15IQTFt9+x45kz57lq2fUaq6N6W3a1L1mYTWfr8n09a3W341pj63f3dmb5fnK9KY5TtNY7cfW+cqoWd5HLNUsa5vVtlbz/l6rHFuYzP0A68VMx9htrV3bWntKkqOS/GiSW6YbciEj7/tcxXLVBgAAAAAwFLN+eFqSpLV2eWvtE621q5I8Kcm5rbVPLdS+qo5LcvskF8yqRgAAAACA1WqmQzFU1V2T/FS68XI3JHlUkgf20+bbPDjJo5O8K8nuJMcmeU6SzyU5fZb1AgAAAACsRrMeY/ebSX45yfOTfCvJOUnu0Vr7xEibzye5RbqHpN0syd4k/5jk2a21K2daLQAAAADAKjTTYLe1dkGSu+ynzceT3Hc2FQEAAAAADM+KjLELAAAAAMD0BLsAAAAAAAMj2AUAAAAAGJhZPzwN1q25uWTbtqUts3lzsmXL8tQzaufOZNeupS2zd+/y1MK+LfU8WqvHaa3th2l+B5PZXSOmtZqve8zWtOf4Us3yd30tfiZgsrV23wGjDuT/Z0u9b1ur97ywkgS7MAMbNya7dyfbty9+mbm5brmtW5evrnm7diVnnZW0trTlNmxYnnqYbJrzKFl7x2kt7odpfgdneY2Yxmq/7jFb0/5/Zhqz+l1fi58J+G5r8b4DRk37/7Np7tvW4j0vrDTBLszApk3dayl27Ej27FmeeiZpLTnhhNltj6Wb5jxai9bqfljq7+CsrxFLNYTrHrO1Fv8/sxY/E/Cd1up9B4ya5v9n0963rbV7XlhpxtgFAAAAABgYwS4AAAAAwMAIdgEAAAAABkawCwAAAAAwMIJdAAAAAICBEewCAAAAAAyMYBcAAAAAYGAOXekCgINn585k166lL7d378GvhWGbm0u2bVt8e+fQMExzjZj1sV3quTcEmzcnW7Ysvv1aPE5DuEasxc+0VEM494D1wTW5M4v7ogPZd2vxOE2zz5d6rwcHk2AX1pBdu5KzzkpaW/qyGzYc/HoYpo0bk927k+3bl7acc2j1m/YaMatjO+25t5rNzXWfa+vWxS+zVo/Tar5GrMXPNI3Vfu4B64NrcmeW90XT7Lu1eJym+UzT3OvBwSTYhTWmteSEE1a6CoZs06buxdq0mq8Ra/Hc27Ej2bNn6cs5TrO1Fj/TtFbzuQesD67JndW+H1Z7fdOY5jNNe68HB4sxdgEAAAAABkawCwAAAAAwMIJdAAAAAICBEewCAAAAAAyMYBcAAAAAYGAEuwAAAAAAAyPYBQAAAAAYmENXugAOzLnndu93v/vK1gEATDY3l2zbtvj2e/cuXy0AAEPjXgoWJtgduL17kx07BLsAsBpt3Jjs3p1s37605TZsWJ56AACGxL0U7JtgFwBgmWza1L0AAFg691Kwb8bYBQAAAAAYGMEuAAAAAMDACHYBAAAAAAZGsAsAAAAAMDCCXQAAAACAgRHsAgAAAAAMjGAXAAAAAGBgDl3pAjhwX/1qsm3b8m9n797l38aszc0tfd/Ncj8stb61eIxmbZpzYvPmZMuW5amHleV8AAAAYLUS7A7c5s3J+9+fbN8+m+1t2DCb7czCxo3J7t3T7btZ7Idp61tLx2jWptnnc3Pdclu3Ll9drAznAwAAAKuZYHfgtmzpwt3v//6VrmR4Nm3qXqvVaq9vLZpmn+/YkezZszz1sLKcDwAAAKxmxtgFAAAAABgYwS4AAAAAwMAIdgEAAAAABmbmwW5V/XxVfbyqvlFVn6mqp01oU1X17Kr6fFV9rao+UFV3nHWtAAAAAACr0UyD3aq6R5IzkuxI8pAkpyXZWlW/O9b0D5I8N8nWvt1VSc6uqlvNsFwAAAAAgFXp0Blv7+QkH2ytPb7/+T1VdXiSk6vq9a21a6rqhumC3Re31l6bJFX1oSQXJ3lKkufMuGYAAAAAgFVl1kMx3DHJ2WPT3pPk8CR363++e5INSf52vkFrbS7JtiQPmkGNAAAAAACr2qyD3RsmuWZs2jf69y39+7FJrk3y6bF2O/t5AAAAAADr2qyHYtiV5C5j036if795/354kqtaa9eOtbs8yWFVdYPW2ng4zCo2N5ds27a0ZTZvTrZs2X87gNVmmmve3r2z2da02wEAAFbOzp3Jrl3TLStfWdtmHeyekuQNVfWEJG9PF+o+vZ83GuS2CcvWQvOq6olJnpgkRx999EErlgO3cWOye3eyffvil5mb65bbunX56gJYDtNc8+Zt2DCbbS11OwAAwMratSs566ykTUrL9kG+svbNOtg9LcmPJXlDklOTXJ3kmUlek+Syvs3lSW5SVYeM9dq9WZKrW2vfHF9pa+3Ufn05/vjjl3ias5w2bepeS7FjR7Jnz/LUA7CcprnmDWFbAADAymotOeGEpS0jX1n7ZjrGbmvt2tbaU5IcleRHk9wyyXn97Pn3TyU5JMnmscWP7ecBAAAAAKxrs354WpKktXZ5a+0TrbWrkjwpybmttfnQ9twkVyb5xfn2VXVYkockOWvmxQIAAAAArDIzHYqhqu6a5KeSXJhkQ5JHJXlgPy1J0lr7elW9JMlzq+rydL10n5YuhH7NLOsFAAAAAFiNZj3G7jeT/HKS5yf5VpJzktyjtfaJsXYvSRfkPivJEUnOT3L/1tplAQAAAABY52Ya7LbWLkhyl0W0a0le1L8AAAAAABixImPsAgAAAAAwPcEuAAAAAMDAzHqMXViUublk27bFt9+7d/lqAQAAAIDVRrDLqrNxY7J7d7J9+9KW27BheeoBAAAAgNVGsMuqs2lT9wIAAAAAJjPGLgAAAADAwAh2AQAAAAAGRrALAAAAADAwgl0AAAAAgIER7AIAAAAADIxgFwAAAABgYAS7AAAAAAADc+hKFwAwdHNzybZti2+/d+/y1QIAAMDsLPXfg0myeXOyZcvy1DNutdfHgRHsAhyAjRuT3buT7duXttyGDctTDwAAALMxzb8H5+a65bZuXb665q32+jhwgl2AA7BpU/cCAABgfZnm34M7diR79ixPPeNWe30cOGPsAgAAAAAMzEEJdqvq+lV19MFYFwAAAAAA+7bfYLeqnlxV/1lVX62qD1fVr01odqcknzn45QEAAAAAMG6fwW5VnZjkNUnOS/KCJLuTnF5Vb6+qG82gPgAAAAAAxuyvx+7vJ/mT1tqvttb+pLX28LRsMwEAABjtSURBVCQPSPJTSd5XVUcse4UAAAAAAHyH/QW7t0+yfXRCa+29Se6a5KZJPlRVP7RMtQEAAAAAMMH+gt0rkhw5PrG1dnGSuyfZk+TcJHc56JUBAAAAADDR/oLdC5L8/KQZrbXLk9w3yflJXn2Q6wIAAAAAYAH7C3bflOQHq+rmk2a21r6W5KFJ/k+Szx3k2gAAAAAAmODQfc1srf1dkr/bT5trkzzxYBYFAAAAAMDC9hnsjquq6yU5ov9xb2vtWwe/JAAAAAAA9mV/QzEkSarqoVV1dpKrknyxf11VVf9UVT+3nAUCAAAAAPCd9ttjt6pel+S3kvxnklOSfLafdZskD05yZlW9obX2lGWrEgAAAACAb9tnsFtVJ6UbP/dJSU4dH3qhqp6e5AlJXltVO1prb1yuQgEAAAAA6OxvKIbfSvKa1topk8bTbZ1Tk7wuXfgLAAAAAMAy21+w+yNJ3rWI9WxLctyBlwMAAAAAwP7sL9i9NskNFrGeG/RtAQAAAABYZvsLdi9IcuIi1nNi3xYAAAAAgGW2z4enJXlFkjOr6tIkL26tXTU6s6punORZSX4tyUOXp0QAAAAAAEbtM9htrW2rqmcleVGS36qqf07y2X72bZLcK8lNkzyrtfYPy1koAAAAAACd/fXYTWtta1WdneQpSX46yYP7WZcmeWeS17bWPrp8JQIAAAAAMGq/wW6StNYuSPLYZa4FAAAAAIBFWFSwO66qvj/JxiRfaK1dcnBLAgAAAABgX663lMZV9VtV9fl04+x+OMlnq+qSqnrSEtZxYlV9tKquqqpLq+qNVXXrsTYXV1Ube31xKbUCAAAAAKxVi+6xW1UnJ3lekr9IckaSLyW5RZJHJnl1VR3ZWnvhftbx0CRvTfK6JP8rXa/fP0ryrqo6vrX2rZHmb0nympGfr1lsrQAAAAAAa9lShmJ4cpI/bq09d2z6P1bVZf38fQa7SX4lyUdba0+Zn1BVVyY5M8ntk+wcafuF1tp5S6gPAAAAAGBdWMpQDDdK8oEF5v1LkhsuYh3XT3LF2LSv9O+1hFoAAAAAANatpQS7f5/kEQvMe2SSdy1iHacl+emqekxVbaiq26UbiuF9rbVPjrX9jaq6pqquqKq3V9VtllArAAAAAMCatZShGM5K8tKqOiZdyDs/xu7Dk/xIkmdU1QnzjVtr28dX0Fr7h6o6Kd04vX/VTz43yUPHmp6Z5LwklyTZkm5s33Oq6g6ttfEevwAAAAAA68pSgt039+/fl+SB+5ifJC3JIeMNqureSU5J8qp0QfEtkzw/yTuq6n6ttWuTpLX2OyOLnVNV5ya5MMljk7xywnqfmOSJSXL00Ucv4SMBAAAAAAzPUoLdHzgI23t5kne21p45P6GqLkzyqSQPS3LGpIVaa/9WVRcludMC809NcmqSHH/88e0g1AkAAAAAsGotOthtrX32IGzv2CRvHVvvRVX1tSQ/tJgyDkINAAAAAACDtpQeu99WVYcleVy6oPaLSd64yOD3sxnrdVtVW5LcKMnF+9jecUlun+TPpqkXAAAAAGAt2WewW1UvT/KQ1trtRqbdJMlHktw2yeVJbprk6VX1E621/9jP9k5J8oqq2p3rxtg9OV2ou71f/4OTPDrJu5LsThcePyfJ55KcvrSPBwAAAACw9lxvP/PvneRNY9N+P8ntkjyhtXZkklunC2afu4jtvTrJk5PcP8mZSV6a7qFo922tzfVtPp/kFukekvaeJM9L8k9Jfqq1duUitgEAAAAAsKbtbyiGY5JcMDbtkUk+2Vo7LUlaa1/ue/a+YH8ba621JG/oXwu1+XiS++5vXQAAAAAA69X+euwemuTr8z9U1c2TbEnyz2PtLk5yq4NaGQAAAAAAE+0v2P2PJPca+fnn+vd3j7W7RZL/Pkg1AQAAAACwD/sbiuG1Sf68qm6a5LIkT03ymXRj3456QJJ/O/jlAQAAAAAwbp/Bbmvt9KramO6BZzdL8tEkT26tfXO+TVUdleRhWcQYuwAAAAAAHLj99dhNa+3FSV68j/lfjvF1AQAAAABmZn9j7AIAAAAAsMoIdgEAAAAABkawCwAAAAAwMIJdAAAAAICBEewCAAAAAAyMYBcAAAAAYGAEuwAAAAAAAyPYBQAAAAAYmENXugAAAAAAWC/m5pJt2xbffu/e5auFYRPsAgAAAMAMbNyY7N6dbN++tOU2bFieehg2wS4AAAAAzMCmTd0LDgZj7AIAAAAADIxgFwAAAABgYAS7AAAAAAADI9gFAAAAABgYwS4AAAAAwMAIdgEAAAAABkawCwAAAAAwMIJdAAAAAICBEewCAAAAAAyMYBcAAAAAYGAEuwAAAAAAAyPYBQAAAAAYGMEuAAAAAMDACHYBAAAAAAZGsAsAAAAAMDCCXQAAAACAgRHsAgAAAAAMjGAXAAAAAGBgBLsAAAAAAAMj2AUAAAAAGBjBLgAAAADAwAh2AQAAAAAGRrALAAAAADAwMw92q+rEqvpoVV1VVZdW1Rur6tZjbaqqnl1Vn6+qr1XVB6rqjrOuFQAAAABgNZppsFtVD03y1iTnJnlYkmcmuWeSd1XVaC1/kOS5SbYmeUiSq5KcXVW3mmW9AAAAAACr0aEz3t6vJPloa+0p8xOq6sokZya5fZKdVXXDdMHui1trr+3bfCjJxUmekuQ5M64ZAAAAAGBVmfVQDNdPcsXYtK/079W/3z3JhiR/O9+gtTaXZFuSBy13gQAAAAAAq92sg93Tkvx0VT2mqjZU1e2S/FGS97XWPtm3OTbJtUk+Pbbszn4eAAAAAMC6NtNgt7X2D0lOSnJqup67FyU5JMkjRpodnuSq1tq1Y4tfnuSwqrrBDEoFAAAAAFi1Zv3wtHsnOSXJq5LcO8mJSW6e5B1VdchI0zZp8YXmVdUTq+r8qjr/y1/+8kGuGgAAAABgdZn1w9NenuSdrbVnzk+oqguTfCrJw5Kcka5n7k2q6pCxXrs3S3J1a+2b4yttrZ2arhdwjj/++EmhMAAAAADAmjHrMXaPTXLh6ITW2kVJvpbkh/pJn0o3PMPmCct+arkLBAAAAABY7WYd7H42yZ1GJ1TVliQ3SnJxP+ncJFcm+cWRNocleUiSs2ZSJQAAAADAKjbroRhOSfKKqtqdLqS9ZZKT04W625Oktfb1qnpJkudW1eXpeuk+LV0I/ZoZ1wsAAAAAsOrMOth9dZJrkvxWkt9M8pUkH0zyrNba3Ei7l6QLcp+V5Igk5ye5f2vtstmWCwAAAACw+sw02G2ttSRv6F/7a/ei/gUAAAAAwIhZj7ELAAAAAMABEuwCAAAAAAzMrMfYBQAAAABWqbm5ZNu2pS2zeXOyZcvi2+/cmezatbRtzDviiOTud59u2bVGsAsAAAAAZOPGZPfuZPv2xS8zN9ctt3Xr4pfZtSs566yktaXXeOtbC3bnCXYBAAAAgGza1L2WYseOZM+epW+rteSEE5a+3CWXLH2ZtcoYuwAAAAAAAyPYBQAAAAAYGMEuAAAAAMDACHYBAAAAAAZGsAsAAAAAMDCCXQAAAACAgRHsAgAAAAAMjGAXAAAAAGBgBLsAAAAAAAMj2AUAAAAAGBjBLgAAAADAwAh2AQAAAAAGRrALAAAAADAwgl0AAAAAgIER7AIAAAAADIxgFwAAAABgYAS7AAAAAAADI9gFAAAAABgYwS4AAAAAwMAIdgEAAAAABkawCwAAAAAwMIJdAAAAAICBEewCAAAAAAyMYBcAAAAAYGAEuwAAAAAAAyPYBQAAAAAYGMEuAAAAAMDACHYBAAAAAAbm0JUuAAAAOLjm5pJt25a2zN69y1MLALD2LfXew33HwSHYBQCANWTjxmT37mT79qUvu2HDwa8HAFjbpr33cN9x4AS7AACwhmza1L0AAGbBvcfKMcYuAAAAAMDACHYBAAAAAAZmpsFuVb2/qtoCr7v1bS6eMO+Ls6wTAAAAAGA1m/UYu09KMj408guT/HiSj4xMe0uS14z8fM0y1wUAAAAAMBgzDXZba58c/bmqbpDk+CR/01r7n5FZX2itnTfL2gAAAAAAhmKlx9j92SSHJ3nrCtcBAAAAADAYKx3snpjk0iTnjE3/jaq6pqquqKq3V9VtVqA2AAAAAIBVadZj7H5bVR2W5CFJTm2ttZFZZyY5L8klSbYkeV6Sc6rqDq21K2ZfKQAAAADA6rJiwW66UPd7MzYMQ2vtd0Z+PKeqzk1yYZLHJnnlpBVV1ROTPDFJjj766GUpFgAAAABgtVjJoRhOTLKrtXb+vhq11v4tyUVJ7rSPNqe21o5vrR1/1FFHHeQyAQAAAABWlxUJdqvqpkkelKU9NK3tvwkAAAAAwNq3Uj12H57ke7KIYLeqjkty+yQXLHdRAAAAAABDsFJj7J6Y5GOttZ2jE6vqwUkeneRdSXYnOTbJc5J8LsnpM64RAAAAAGBVmnmwW1VHJrlvkudOmP35JLdI95C0myXZm+Qfkzy7tXblzIoEAAAAAFjFZh7sttb2JLn+AvM+ni70BQAAAABgASs1xi4AAAAAAFMS7AIAAAAADIxgFwAAAABgYGY+xi4H3xFHJJdcstJVAAAAAMDyOuKIla5g9RDsrgG/9EsrXQEAAAAAMEuGYgAAAAAAGBjBLgAAAADAwAh2AQAAAAAGRrALAAAAADAwgl0AAAAAgIER7AIAAAAADIxgFwAAAABgYAS7AAAAAAADI9gFAAAAABgYwS4AAAAAwMAIdgEAAAAABqZaaytdw0FVVV9O8tmVrmPGjkyyZ6WLAFYt1whgIa4PwL64RgD74hoBs3Gb1tpRk2asuWB3Paqq81trx690HcDq5BoBLMT1AdgX1whgX1wjYOUZigEAAAAAYGAEuwAAAAAAAyPYXRtOXekCgFXNNQJYiOsDsC+uEcC+uEbACjPGLgAAAADAwOixCwAAAAAwMILdgaqqH66q91bV1VW1u6peWFWHrHRdwGxV1S9W1Tur6tKquqqqLqiqR421qap6dlV9vqq+VlUfqKo7rlTNwMqpqu/rrxWtqr53ZLrrBKxTVXVoVf1BVX26qr5RVZdU1SvG2rhGwDpUVSdW1Uf7e4dLq+qNVXXrsTauD7CCBLsDVFWHJzk7SUvysCQvTPL0JC9YybqAFfG0JFcl+b0kD03yviRvqarfHmnzB0mem2Rrkof07c+uqlvNuFZg5b0s3TVgnOsErF9/meSpSf4kyQPSXQ++NtbGNQLWmap6aJK3Jjk3Xe7wzCT3TPKuqhrNklwfYAUZY3eAqupZSZ6R5DattSv7ac9I8vwkt5qfBqx9VXVka23P2LS3JLlba+0HquqGSS5L8vLW2gv7+TdOcnGSP2utPWfWNQMro6p+OsmZSf44XcB7k9baVa4TsH5V1c8m2Zbkx1prn1ygjWsErENV9bYkt22t3Xlk2kPT3Uv8cGttp+sDrDw9dofpQUnePRbgvi3JjZL8zMqUBKyE8VC3969JbtH/992TbEjytyPLzKX7R9yDlr1AYFXoh2t6Tbpv+YxfN1wnYP36jST/vFCo23ONgPXp+kmuGJv2lf69+nfXB1hhgt1hOjbJp0YntNY+l+Tqfh6wvt09yfw/0I5Ncm2ST4+12RnXC1hPfjPJDZO8bsI81wlYv34yyX9U1Wur6sr++R1njI2h6RoB69NpSX66qh5TVRuq6nZJ/ijJ+0b+GOT6ACtMsDtMh+e6v5SNuryfB6xTVXXfdGNgzYc3hye5qrV27VjTy5McVlU3mGV9wOxV1RFJ/neSp7XWvjmhiesErF+3SnJSkjsmOTHJY5PcOck7qmq+R55rBKxDrbV/SHd9ODVdz92LkhyS5BEjzVwfYIUdutIFMLVJgyPXAtOBdaCqjknyliRnttZOH5m10PVioXnA2vKiJB9urW3fRxvXCVifqn89rLW2N0mq6gtJ/iXJfZK8t2/nGgHrTFXdO8kpSV6V5Kwkt0z3XJ93VNX9RsJc1wdYQYLdYbo8yc0mTL9pJvfkBda4qrp5uhuuzyV59Misy5PcpKoOGftL+s2SXL1A7z1gjaiqH0k3huY9q2r+3uGw/v2mVXVtXCdgPbs8yX/Nh7q9Dya5JskPpwt2XSNgfXp5kne21p45P6GqLkw3LOTDkpwR1wdYcYZiGKZPZWy8mqralOTGGRt7F1j7quqwJO9KcoMkD+4fWDDvU+m+MrV5bLHvGqsbWJNum+7hJx9K94+vy3PdUC2XpHugmusErF87F5heSb7V/7drBKxPxya5cHRCa+2iJF9L8kP9JNcHWGGC3WE6K8kDq+omI9N+Od0F9l9WpiRgJVTVoUn+Ll1486DW2pfGmpyb5MokvziyzGFJHpLuWgKsbR9Mcu+x19Z+3glJXhbXCVjP3pXkR6vqyJFp90z3B6GP9T+7RsD69NkkdxqdUFVbktwoycX9JNcHWGGGYhimU5I8NckZVbU1yQ+mG+vmT1trV65kYcDMvT5dOPM7SW5eVXcdmfevrbWvV9VLkjy3qi5P95fzp6X7w95rZl4tMFOttT1J3j86rR+PO0nOaa1d1U9znYD16dR0/67YVlV/nOQm6f74c3Zr7YNJ4l4C1q1TkryiqnbnujF2T04X6m5PXB9gNRDsDlBr7fLqnnz/2iTb0o2r+4p04S6wvjygf3/VhHk/kO7G6yXpbq6eleSIJOcnuX9r7bJZFAgMgusErEOttSur6j5JXp3kbenG1j0zye+NNXWNgPXn1emuCb+V5DfT5Q4fTPKssaHfXB9gBVVrHlIIAAAAADAkxtgFAAAAABgYwS4AAAAAwMAIdgEAAAAABkawCwAAAAAwMIJdAAAAAICBEewCAAAAAAyMYBcAgFWhqtoiXvf6/9u7/1C/6jqO48+XjqxQGpLKKEJLhljI+iNxFpT4xwgH5fzB0D80MENzwkozLFZIqYkNRIXYZqg4dkV0Rv7A0tkfQk5WE43VluJNRrKktRZas7V3f5zPF758d933jnm993vv8wGHc87nfD7nvM/96/Lm/X2fJONJbp/ueN8LSU5u77X0MNctTPLDJPOnKjZJkiTNbKmq6Y5BkiRJIslZfacfAjYBPwIe7xvfBnwK+HtVvf4+hjclkhwDfBb4U1XtOYx1S4FfAqdU1fgUhSdJkqQZbN50ByBJkiQBVNXzveMkx7bDV/vHm63vX1RTq6r2AYPvJ0mSJA1lKwZJkiSNlMFWDEnuTbIlyXlJtiV5O8njSY5PcmqSZ5O81eacMXCvo5J8N8krSfYl2ZHksknEUEm+leSOJLuT7ElyZ5IPDMxblOSZFtM/kqxPclLf9YNaMfTeL8nKJDvburFe24UkX6Kr1gV4ra0fb9fmJ1mX5K9J/pPk9SRrD/+vLEmSpJnOxK4kSZJmg08ANwHfB64EzgbWAGNtu5Du12pjSdK37s62Zg1wHrAR+Pkke95+G/g4cCldy4grgR/3LiY5AfgN8GHgEmAF8EXg14MJ4AlcDJzb7nkDsBS4uV37PXBdO14GLAbOb+ergS8AK4ElwI2AvdckSZJmIVsxSJIkaTY4HlhcVa8CtMrc64HLqur+Nha6fr2nAX9McipwFfC1qrqv3efpJAuAHwCPDXnmv4CLquoA8GTrl/u9JLdU1W66xC/Akqra22LYAWwGLgA2HOLe/wW+WlX727rTgeXA1VW1N8n2Nm/rQI/dM4G7q+rBvrEHhryHJEmSRpAVu5IkSZoNxntJ3eaVtt80wdjH2v5c4ACwMcm83gY8AyxKcvSQZ/6iJXV7HqH76Ntn2vmZwK96SV2AqnoBGKerqj2UZ3tJ3WYbcOIkKn1fBK5PcnWShUPmSpIkaYSZ2JUkSdJssGfg/J0JxntjH2z7jwJHA/+kq5DtbffS/bJtwZBn/u1dzhf07XdNsG4XXYXxoUz0PgGGJXavAR4FVgHbk/w5yfIhayRJkjSCbMUgSZKkuWo3sB/4PF3l7qDBxO2gE9/l/I2+/eAcgJOA300yxsNSVXuAa4FrWzuK7wDrk7xUVdum4pmSJEmaHlbsSpIkaa7aRFex+5Gq2jLB9s6Q9V9J0v//9DLg38Af2vlmYEmS43oTknwOOBl47ghjH6w+PkhVvUTXZ/gour7CkiRJmkWs2JUkSdKcVFXbk/wMGEtyG7CFLlH6aWBhVV0x5BbHAQ8lWdvWrALuah9OA1hN93G2p5L8BDgWuBV4GXj4CMPvfTztG0nGgLer6uUkzwEb6ZLLBXwdeAt44QifJ0mSpBnGxK4kSZLmsm8CO+gSoDcBe+k+VHbPJNb+FPgksIGuKnYdcGPvYlW9meScNm8DXZXtE8DKSVQDH1JV/SXJdXRtF1YAO+kqgX8LXN6O/wdsBb5cVTuP5HmSJEmaeVJV0x2DJEmSNFKSFLCiqu6a7lgkSZI0N9ljV5IkSZIkSZJGjIldSZIkSZIkSRoxtmKQJEmSJEmSpBFjxa4kSZIkSZIkjRgTu5IkSZIkSZI0YkzsSpIkSZIkSdKIMbErSZIkSZIkSSPGxK4kSZIkSZIkjRgTu5IkSZIkSZI0Yv4PWAg2hJCJvrAAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 1728x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_seq('GMA000', 'blue')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1728x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_seq('SJ110834', 'red')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1728x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_seq('FN230635', 'orange')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1728x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plot_seq('WJ151131', 'purple')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Dựng mô hình \n",
    "\n",
    "Nhi dùng method GaussianHMM của thư viện hmmlearn, với algorithm là Viterbi để luyện một mô hình HMM với đầu ra phân phối Gaussian; đặc biệt việc huấn luyện thực hiện trên nhiều đơn vị quan sát (chuỗi) với độ dài khác nhau.\n",
    "\n",
    "Để thực hiện điều này, đầu vào cần có X là 1 array 2 chiều, và cần 1 array lengths chứa độ dài của từng chuỗi (ta đã có ở trên)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[96],\n",
       "       [98],\n",
       "       [97],\n",
       "       ...,\n",
       "       [93],\n",
       "       [93],\n",
       "       [94]], dtype=int64)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Dữ liệu đầu vào có dạng thế này:\n",
    "\n",
    "X = np.concatenate([seq_db[p] for p in seq_db.keys()])\n",
    "\n",
    "np.atleast_2d(X).T"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[14, 19, 10, 16, 29, 48, 18, 22, 55, 80, 40, 30, 93, 23, 43, 19, 19, 42, 77, 51, 18, 32, 19, 18, 14, 13, 15, 8, 59, 54, 26, 17, 87, 16, 65]\n"
     ]
    }
   ],
   "source": [
    "print(lengths)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Một tham số quan trọng mang tính chủ quan trong mô hình HMM, đó là số hidden states (n components) mà ta muốn giả định; Nhi sẽ khảo sát score của mô hình với giá trị n từ 2 đến 12:\n",
    "\n",
    "Một tham số khác, đó là cách tính covariance, có 2 phương pháp là Diag và full (giống như trong algorithm Gaussian Mixture clustering), tuy nhiên sẽ không có khác biệt nếu dữ liệu chỉ gồm 1 chuỗi duy nhất (SpO2), trong trường hợp có nhiều chuỗi song song, phương pháp 'Full' có thể cho ra kết quả khác với 'Diag'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5, 0, 'n_components')"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "n_components = np.arange(2, 13)\n",
    "\n",
    "models1 = [GaussianHMM(n_components=n, \n",
    "                    covariance_type=\"diag\",\n",
    "                    n_iter=2000).fit(np.atleast_2d(X).T, lengths)\n",
    "           for n in n_components]\n",
    "          \n",
    "models2 = [GaussianHMM(n_components=n, \n",
    "                    covariance_type=\"full\",\n",
    "                    n_iter=2000).fit(np.atleast_2d(X).T, lengths)\n",
    "           for n in n_components]\n",
    "         \n",
    "plt.figure(figsize=(10,6))\n",
    "\n",
    "plt.plot(n_components, \n",
    "         [m.score_samples(np.atleast_2d(X).T, lengths)[0] for m in models1], label='Diag covar',\n",
    "         c = 'red'\n",
    "        )\n",
    "          \n",
    "plt.plot(n_components, \n",
    "         [m.score_samples(np.atleast_2d(X).T, lengths)[0] for m in models2], label='Full covar',\n",
    "         c = 'blue'\n",
    "        )\n",
    "\n",
    "plt.legend(loc='best')\n",
    "plt.xlabel('n_components')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Có vẻ như một mô hình với 5 hidden states là phù hợp; ta sẽ dựng mô hình này"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "GaussianHMM(algorithm='viterbi', covariance_type='diag', covars_prior=0.01,\n",
       "            covars_weight=1, init_params='stmc', means_prior=0, means_weight=0,\n",
       "            min_covar=0.001, n_components=5, n_iter=5000, params='stmc',\n",
       "            random_state=None, startprob_prior=1.0, tol=0.01,\n",
       "            transmat_prior=1.0, verbose=False)"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = GaussianHMM(n_components=5, \n",
    "               covariance_type=\"diag\",\n",
    "               n_iter=5000)\n",
    "\n",
    "model.fit(np.atleast_2d(X).T, lengths)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Diễn giải mô hình\n",
    "\n",
    "Đầu tiên, Nhi sẽ khảo sát đặc tính phân bố của 5 trạng thái ẩn mà mô hình học được từ dữ liệu:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Trung bình ± SD tương ứng 5 hidden states\n",
      "0th hidden state\n",
      "97.87891639420545 ± 0.8128419110991656\n",
      "----------------------------------------\n",
      "1th hidden state\n",
      "90.80917007433679 ± 2.511330051159424\n",
      "----------------------------------------\n",
      "2th hidden state\n",
      "93.93331886555568 ± 1.9248274230272753\n",
      "----------------------------------------\n",
      "3th hidden state\n",
      "82.41847411162121 ± 8.91257013052585\n",
      "----------------------------------------\n",
      "4th hidden state\n",
      "96.1279816142075 ± 1.2372715933339342\n",
      "----------------------------------------\n"
     ]
    }
   ],
   "source": [
    "print(\"Trung bình ± SD tương ứng 5 hidden states\")\n",
    "\n",
    "for i in range(model.n_components):\n",
    "    print(f'{i}th hidden state')\n",
    "    print(f'{float(model.means_[i])} ± {float(np.diag(model.covars_[i])**0.5)}')\n",
    "    print('-'*40)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Hình ảnh của 5 phân bố Gaussian tương ứng  hidden states của mô hình HMMG:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "tfd = tfp.distributions\n",
    "\n",
    "mu = model.means_.flatten()\n",
    "sd = model.covars_.flatten()**0.5\n",
    "\n",
    "plt.figure(figsize=(12,7))\n",
    "\n",
    "for i in [0,1,2,3,4]:\n",
    "    loc = tf.constant(mu[i])\n",
    "    scale =  tf.constant(sd[i])\n",
    "    dist = tfd.Normal(loc = loc, scale = scale)\n",
    "    samples = dist.sample(1000).numpy()\n",
    "    \n",
    "    sns.kdeplot(pd.Series(samples, name = str(f'State{i}: mu={np.round(loc,2)},sigma={np.round(scale,2)}')),shade = True, alpha = 0.7)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Ma trận chuyển tiếp giữa 5 trạng thái ẩn được xuất ra từ mô hình như sau:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>97.88</th>\n",
       "      <th>90.81</th>\n",
       "      <th>93.93</th>\n",
       "      <th>82.42</th>\n",
       "      <th>96.13</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>97.88</th>\n",
       "      <td>8.267844e-01</td>\n",
       "      <td>3.330622e-30</td>\n",
       "      <td>0.124441</td>\n",
       "      <td>1.402555e-52</td>\n",
       "      <td>4.877486e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>90.81</th>\n",
       "      <td>3.813857e-06</td>\n",
       "      <td>8.918647e-01</td>\n",
       "      <td>0.050722</td>\n",
       "      <td>4.964190e-02</td>\n",
       "      <td>7.767641e-03</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>93.93</th>\n",
       "      <td>2.124992e-02</td>\n",
       "      <td>1.828578e-02</td>\n",
       "      <td>0.908204</td>\n",
       "      <td>8.361198e-03</td>\n",
       "      <td>4.389947e-02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>82.42</th>\n",
       "      <td>1.307522e-30</td>\n",
       "      <td>3.821833e-02</td>\n",
       "      <td>0.118668</td>\n",
       "      <td>8.431137e-01</td>\n",
       "      <td>1.831411e-08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>96.13</th>\n",
       "      <td>7.587430e-04</td>\n",
       "      <td>3.966586e-08</td>\n",
       "      <td>0.084482</td>\n",
       "      <td>1.934286e-06</td>\n",
       "      <td>9.147577e-01</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              97.88         90.81     93.93         82.42         96.13\n",
       "97.88  8.267844e-01  3.330622e-30  0.124441  1.402555e-52  4.877486e-02\n",
       "90.81  3.813857e-06  8.918647e-01  0.050722  4.964190e-02  7.767641e-03\n",
       "93.93  2.124992e-02  1.828578e-02  0.908204  8.361198e-03  4.389947e-02\n",
       "82.42  1.307522e-30  3.821833e-02  0.118668  8.431137e-01  1.831411e-08\n",
       "96.13  7.587430e-04  3.966586e-08  0.084482  1.934286e-06  9.147577e-01"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tm_df = pd.DataFrame(model.transmat_, \n",
    "                     columns=np.round(model.means_.flatten(),2),\n",
    "                     index=np.round(model.means_.flatten(),2))\n",
    "\n",
    "tm_df"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x2cae9b0d388>"
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 720x576 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.rcParams['figure.figsize'] = (10,8)\n",
    "\n",
    "sns.heatmap(tm_df, cmap = 'viridis', annot=True, fmt=\".3f\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Từ ma trận chuyển tiếp, ta có thể vẽ thành mô hình mạng (network) như sau:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/svg+xml": [
       "<?xml version=\"1.0\" encoding=\"UTF-8\" standalone=\"no\"?>\r\n",
       "<!DOCTYPE svg PUBLIC \"-//W3C//DTD SVG 1.1//EN\"\r\n",
       " \"http://www.w3.org/Graphics/SVG/1.1/DTD/svg11.dtd\">\r\n",
       "<!-- Generated by graphviz version 2.38.0 (20140413.2041)\r\n",
       " -->\r\n",
       "<!-- Title: %3 Pages: 1 -->\r\n",
       "<svg width=\"315pt\" height=\"245pt\"\r\n",
       " viewBox=\"0.00 0.00 315.47 245.46\" xmlns=\"http://www.w3.org/2000/svg\" xmlns:xlink=\"http://www.w3.org/1999/xlink\">\r\n",
       "<g id=\"graph0\" class=\"graph\" transform=\"scale(1 1) rotate(0) translate(4 241.459)\">\r\n",
       "<title>%3</title>\r\n",
       "<polygon fill=\"white\" stroke=\"none\" points=\"-4,4 -4,-241.459 311.472,-241.459 311.472,4 -4,4\"/>\r\n",
       "<!-- 97.88 -->\r\n",
       "<g id=\"node1\" class=\"node\"><title>97.88</title>\r\n",
       "<ellipse fill=\"none\" stroke=\"black\" cx=\"154.74\" cy=\"-18\" rx=\"30.5947\" ry=\"18\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"154.74\" y=\"-14.3\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">97.88</text>\r\n",
       "</g>\r\n",
       "<!-- 97.88&#45;&gt;97.88 -->\r\n",
       "<g id=\"edge1\" class=\"edge\"><title>97.88&#45;&gt;97.88</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M183.164,-24.7236C194.153,-25.0225 203.287,-22.7813 203.287,-18 203.287,-14.8623 199.354,-12.8185 193.543,-11.8686\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"193.347,-8.35181 183.164,-11.2764 192.948,-15.3404 193.347,-8.35181\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"218.787\" y=\"-14.3\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">0.827</text>\r\n",
       "</g>\r\n",
       "<!-- 90.81 -->\r\n",
       "<g id=\"node2\" class=\"node\"><title>90.81</title>\r\n",
       "<ellipse fill=\"none\" stroke=\"black\" cx=\"227.924\" cy=\"-118.73\" rx=\"30.5947\" ry=\"18\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"227.924\" y=\"-115.03\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">90.81</text>\r\n",
       "</g>\r\n",
       "<!-- 97.88&#45;&gt;90.81 -->\r\n",
       "<g id=\"edge2\" class=\"edge\"><title>97.88&#45;&gt;90.81</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M161.53,-35.7757C171.318,-52.592 189.182,-77.7774 204.001,-96.023\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"201.455,-98.4355 210.555,-103.863 206.826,-93.9457 201.455,-98.4355\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"173.765\" y=\"-69.6993\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">0.0</text>\r\n",
       "</g>\r\n",
       "<!-- 93.93 -->\r\n",
       "<g id=\"node3\" class=\"node\"><title>93.93</title>\r\n",
       "<ellipse fill=\"none\" stroke=\"black\" cx=\"36.3255\" cy=\"-180.984\" rx=\"30.5947\" ry=\"18\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"36.3255\" y=\"-177.284\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">93.93</text>\r\n",
       "</g>\r\n",
       "<!-- 97.88&#45;&gt;93.93 -->\r\n",
       "<g id=\"edge3\" class=\"edge\"><title>97.88&#45;&gt;93.93</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M138.281,-33.1652C114.434,-60.9677 71.2576,-120.106 49.4214,-154.602\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"46.3771,-152.871 44.1077,-163.22 52.3355,-156.545 46.3771,-152.871\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"78.3514\" y=\"-97.6837\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">0.124</text>\r\n",
       "</g>\r\n",
       "<!-- 82.42 -->\r\n",
       "<g id=\"node4\" class=\"node\"><title>82.42</title>\r\n",
       "<ellipse fill=\"none\" stroke=\"black\" cx=\"154.74\" cy=\"-219.459\" rx=\"30.5947\" ry=\"18\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"154.74\" y=\"-215.759\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">82.42</text>\r\n",
       "</g>\r\n",
       "<!-- 97.88&#45;&gt;82.42 -->\r\n",
       "<g id=\"edge4\" class=\"edge\"><title>97.88&#45;&gt;82.42</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M150.693,-35.8781C147.385,-70.9585 147.113,-149.054 149.879,-191.196\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"146.402,-191.634 150.67,-201.331 153.381,-191.089 146.402,-191.634\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"141.286\" y=\"-117.337\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">0.0</text>\r\n",
       "</g>\r\n",
       "<!-- 96.13 -->\r\n",
       "<g id=\"node5\" class=\"node\"><title>96.13</title>\r\n",
       "<ellipse fill=\"none\" stroke=\"black\" cx=\"36.3255\" cy=\"-56.4753\" rx=\"30.5947\" ry=\"18\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"36.3255\" y=\"-52.7753\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">96.13</text>\r\n",
       "</g>\r\n",
       "<!-- 97.88&#45;&gt;96.13 -->\r\n",
       "<g id=\"edge5\" class=\"edge\"><title>97.88&#45;&gt;96.13</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M124.566,-21.6434C107.024,-26.1923 84.9736,-33.2666 67.2976,-39.9799\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"66.002,-36.7285 57.9841,-43.654 68.5709,-43.2401 66.002,-36.7285\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"80.432\" y=\"-19.6116\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">0.049</text>\r\n",
       "</g>\r\n",
       "<!-- 90.81&#45;&gt;97.88 -->\r\n",
       "<g id=\"edge6\" class=\"edge\"><title>90.81&#45;&gt;97.88</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M221.135,-100.954C211.347,-84.1375 193.482,-58.9521 178.664,-40.7066\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"181.209,-38.2941 172.11,-32.8668 175.839,-42.7838 181.209,-38.2941\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"190.899\" y=\"-74.6302\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">0.0</text>\r\n",
       "</g>\r\n",
       "<!-- 90.81&#45;&gt;90.81 -->\r\n",
       "<g id=\"edge7\" class=\"edge\"><title>90.81&#45;&gt;90.81</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M256.348,-125.453C267.338,-125.752 276.472,-123.511 276.472,-118.73 276.472,-115.592 272.538,-113.548 266.728,-112.598\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"266.531,-109.081 256.348,-112.006 266.132,-116.07 266.531,-109.081\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"291.972\" y=\"-115.03\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">0.892</text>\r\n",
       "</g>\r\n",
       "<!-- 90.81&#45;&gt;93.93 -->\r\n",
       "<g id=\"edge8\" class=\"edge\"><title>90.81&#45;&gt;93.93</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M198.122,-123.203C162.976,-132.233 104.8,-151.015 68.6102,-164.924\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"67.0413,-161.78 59.0231,-168.705 69.6099,-168.292 67.0413,-161.78\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"148.866\" y=\"-147.863\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">0.051</text>\r\n",
       "</g>\r\n",
       "<!-- 90.81&#45;&gt;82.42 -->\r\n",
       "<g id=\"edge9\" class=\"edge\"><title>90.81&#45;&gt;82.42</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M210.54,-133.613C197.049,-149.223 178.907,-173.926 166.905,-192.87\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"163.774,-191.282 161.554,-201.642 169.75,-194.927 163.774,-191.282\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"176.223\" y=\"-167.041\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">0.05</text>\r\n",
       "</g>\r\n",
       "<!-- 90.81&#45;&gt;96.13 -->\r\n",
       "<g id=\"edge10\" class=\"edge\"><title>90.81&#45;&gt;96.13</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M205.102,-106.401C173.231,-93.4626 114.964,-74.2382 75.9299,-63.554\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"76.5223,-60.09 65.9589,-60.9049 74.7249,-66.8553 76.5223,-60.09\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"125.016\" y=\"-73.7773\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">0.008</text>\r\n",
       "</g>\r\n",
       "<!-- 93.93&#45;&gt;97.88 -->\r\n",
       "<g id=\"edge11\" class=\"edge\"><title>93.93&#45;&gt;97.88</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M52.7843,-165.819C76.6313,-138.016 119.808,-78.8778 141.644,-44.3817\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"144.689,-46.113 146.958,-35.7641 138.73,-42.439 144.689,-46.113\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"112.714\" y=\"-108.9\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">0.021</text>\r\n",
       "</g>\r\n",
       "<!-- 93.93&#45;&gt;90.81 -->\r\n",
       "<g id=\"edge12\" class=\"edge\"><title>93.93&#45;&gt;90.81</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M66.1279,-176.511C101.274,-167.481 159.45,-148.698 195.64,-134.79\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"197.209,-137.933 205.227,-131.008 194.64,-131.422 197.209,-137.933\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"115.384\" y=\"-159.45\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">0.018</text>\r\n",
       "</g>\r\n",
       "<!-- 93.93&#45;&gt;93.93 -->\r\n",
       "<g id=\"edge13\" class=\"edge\"><title>93.93&#45;&gt;93.93</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M64.7489,-187.707C75.7386,-188.006 84.8728,-185.765 84.8728,-180.984 84.8728,-177.846 80.939,-175.802 75.1286,-174.852\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"74.9321,-171.336 64.7489,-174.26 74.5333,-178.324 74.9321,-171.336\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"100.373\" y=\"-177.284\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">0.908</text>\r\n",
       "</g>\r\n",
       "<!-- 93.93&#45;&gt;82.42 -->\r\n",
       "<g id=\"edge14\" class=\"edge\"><title>93.93&#45;&gt;82.42</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M58.0197,-193.82C73.8439,-200.328 95.7303,-207.737 114.687,-213.125\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"114.003,-216.566 124.571,-215.817 115.842,-209.812 114.003,-216.566\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"70.8531\" y=\"-207.272\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">0.008</text>\r\n",
       "</g>\r\n",
       "<!-- 93.93&#45;&gt;96.13 -->\r\n",
       "<g id=\"edge15\" class=\"edge\"><title>93.93&#45;&gt;96.13</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M41.2261,-163.145C43.3655,-142.619 43.6487,-108.653 42.0759,-84.5681\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"45.5374,-83.9576 41.2226,-74.2812 38.5613,-84.5363 45.5374,-83.9576\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"26.151\" y=\"-127.656\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">0.044</text>\r\n",
       "</g>\r\n",
       "<!-- 82.42&#45;&gt;97.88 -->\r\n",
       "<g id=\"edge16\" class=\"edge\"><title>82.42&#45;&gt;97.88</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M158.787,-201.581C162.095,-166.501 162.367,-88.4046 159.602,-46.2633\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"163.078,-45.8249 158.811,-36.1277 156.099,-46.3698 163.078,-45.8249\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"168.194\" y=\"-127.722\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">0.0</text>\r\n",
       "</g>\r\n",
       "<!-- 82.42&#45;&gt;90.81 -->\r\n",
       "<g id=\"edge17\" class=\"edge\"><title>82.42&#45;&gt;90.81</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M172.125,-204.575C185.616,-188.966 203.758,-164.263 215.759,-145.319\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"218.891,-146.907 221.111,-136.547 212.915,-143.261 218.891,-146.907\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"209.442\" y=\"-178.747\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">0.038</text>\r\n",
       "</g>\r\n",
       "<!-- 82.42&#45;&gt;93.93 -->\r\n",
       "<g id=\"edge18\" class=\"edge\"><title>82.42&#45;&gt;93.93</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M133.046,-206.623C117.222,-200.115 95.3353,-192.706 76.3791,-187.318\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"77.0629,-183.877 66.4945,-184.626 75.2233,-190.631 77.0629,-183.877\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"120.213\" y=\"-185.771\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">0.119</text>\r\n",
       "</g>\r\n",
       "<!-- 82.42&#45;&gt;82.42 -->\r\n",
       "<g id=\"edge19\" class=\"edge\"><title>82.42&#45;&gt;82.42</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M183.164,-226.183C194.153,-226.482 203.287,-224.24 203.287,-219.459 203.287,-216.321 199.354,-214.278 193.543,-213.328\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"193.347,-209.811 183.164,-212.735 192.948,-216.8 193.347,-209.811\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"218.787\" y=\"-215.759\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">0.843</text>\r\n",
       "</g>\r\n",
       "<!-- 82.42&#45;&gt;96.13 -->\r\n",
       "<g id=\"edge20\" class=\"edge\"><title>82.42&#45;&gt;96.13</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M147.14,-202C129.52,-172.401 86.5135,-112.641 59.4528,-79.611\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"62.0414,-77.2505 52.9417,-71.8241 56.6713,-81.7408 62.0414,-77.2505\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"94.2964\" y=\"-144.606\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">0.0</text>\r\n",
       "</g>\r\n",
       "<!-- 96.13&#45;&gt;97.88 -->\r\n",
       "<g id=\"edge21\" class=\"edge\"><title>96.13&#45;&gt;97.88</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M66.4993,-52.8319C84.0413,-48.2829 106.092,-41.2086 123.768,-34.4953\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"125.064,-37.7468 133.082,-30.8212 122.495,-31.2352 125.064,-37.7468\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"102.884\" y=\"-32.4636\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">0.001</text>\r\n",
       "</g>\r\n",
       "<!-- 96.13&#45;&gt;90.81 -->\r\n",
       "<g id=\"edge22\" class=\"edge\"><title>96.13&#45;&gt;90.81</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M59.1475,-68.8043C91.0186,-81.7422 149.286,-100.967 188.32,-111.651\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"187.728,-115.115 198.291,-114.3 189.525,-108.349 187.728,-115.115\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"114.734\" y=\"-94.0275\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">0.0</text>\r\n",
       "</g>\r\n",
       "<!-- 96.13&#45;&gt;93.93 -->\r\n",
       "<g id=\"edge23\" class=\"edge\"><title>96.13&#45;&gt;93.93</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M31.4249,-74.3143C29.2855,-94.8398 29.0023,-128.806 30.5751,-152.891\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"27.1136,-153.501 31.4284,-163.178 34.0897,-152.923 27.1136,-153.501\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"15.5\" y=\"-102.403\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">0.084</text>\r\n",
       "</g>\r\n",
       "<!-- 96.13&#45;&gt;82.42 -->\r\n",
       "<g id=\"edge24\" class=\"edge\"><title>96.13&#45;&gt;82.42</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M43.9257,-73.934C61.5453,-103.533 104.552,-163.293 131.613,-196.323\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"129.024,-198.684 138.124,-204.11 134.394,-194.194 129.024,-198.684\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"78.7693\" y=\"-123.929\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">0.0</text>\r\n",
       "</g>\r\n",
       "<!-- 96.13&#45;&gt;96.13 -->\r\n",
       "<g id=\"edge25\" class=\"edge\"><title>96.13&#45;&gt;96.13</title>\r\n",
       "<path fill=\"none\" stroke=\"black\" d=\"M64.7489,-63.1989C75.7386,-63.4977 84.8728,-61.2565 84.8728,-56.4753 84.8728,-53.3376 80.939,-51.2938 75.1286,-50.3439\"/>\r\n",
       "<polygon fill=\"black\" stroke=\"black\" points=\"74.9321,-46.8271 64.7489,-49.7516 74.5333,-53.8157 74.9321,-46.8271\"/>\r\n",
       "<text text-anchor=\"middle\" x=\"100.373\" y=\"-52.7753\" font-family=\"Times New Roman,serif\" font-size=\"14.00\">0.915</text>\r\n",
       "</g>\r\n",
       "</g>\r\n",
       "</svg>\r\n"
      ],
      "text/plain": [
       "<graphviz.files.Source at 0x2c9332445c8>"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "states = [str(i) for i in tm_df.index]\n",
    "\n",
    "def get_markov_edges(Q):\n",
    "    edges = {}\n",
    "    for col in Q.columns:\n",
    "        for idx in Q.index:\n",
    "            edges[(idx,col)] = Q.loc[idx,col]\n",
    "    return edges\n",
    "\n",
    "edges_wts = get_markov_edges(tm_df)\n",
    "\n",
    "# create graph object\n",
    "G = nx.DiGraph()\n",
    "\n",
    "# edges represent transition probabilities\n",
    "for k, v in edges_wts.items():\n",
    "    tmp_origin, tmp_destination = k[0], k[1]\n",
    "    G.add_edge(tmp_origin, tmp_destination, weight=v, label= np.round(v,3))\n",
    "    \n",
    "pos = nx.spring_layout(G)\n",
    "\n",
    "edge_labels = {(n1,n2):np.round(d['label'],3) for n1,n2,d in G.edges(data=True)}\n",
    "\n",
    "nx.drawing.nx_pydot.write_dot(G, \"tmat.dot\")\n",
    "\n",
    "s = Source.from_file('tmat.dot', engine = 'circo') \n",
    "\n",
    "s"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Một cách tổng quát, ma trận chuyển tiếp cho ta biết xác suất chuyển tiếp từ 1 trạng thái này sang 1 trạng thái khác, vớ tên của trạng thái được đại diện bằng trung bình SpO2 của phân phối Gaussian tương ứng với trạng thái đó; các giá trị trung bình này có một ý nghĩa nhất định về lâm sàng, thí dụ SpO2 = 96 là tốt, 97 là rất tốt, 94 là ranh giới giữa suy hô hấp và bình thường, <90 là suy hô hấp rất nặng, có nguy cơ tử vong cao..."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Áp dụng mô hình cho cá thể\n",
    "\n",
    "Ngoài ứng dụng suy diễn tổng quát, mô hình HMM còn có thể dùng để tiên lượng cho từng cá thể (còn gọi là decode), cho dữ liệu quan sát mô hình sẽ xuất ra trạng thái ẩn và xác suất của mỗi trạng thái;"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(-145.80972574360845,\n",
       " array([1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 4, 4, 4, 4, 4, 4, 4, 4,\n",
       "        4, 4, 4, 4, 2, 2, 2, 2, 2, 0, 0, 0, 0, 2, 4, 4, 4, 4, 4, 4, 4, 4,\n",
       "        4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2]))"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "x1 = np.atleast_2d(seq_db['WJ151131']).T\n",
    "\n",
    "model.decode(x1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 4, 4, 4, 4, 4, 4, 4, 4,\n",
       "       4, 4, 4, 4, 2, 2, 2, 2, 2, 0, 0, 0, 0, 2, 4, 4, 4, 4, 4, 4, 4, 4,\n",
       "       4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2])"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model.predict(x1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0.   , 0.757, 0.187, 0.057, 0.   ],\n",
       "       [0.   , 0.125, 0.753, 0.015, 0.106],\n",
       "       [0.001, 0.038, 0.857, 0.003, 0.101],\n",
       "       [0.   , 0.027, 0.945, 0.001, 0.028],\n",
       "       [0.   , 0.027, 0.972, 0.001, 0.   ],\n",
       "       [0.   , 0.014, 0.985, 0.001, 0.001],\n",
       "       [0.   , 0.005, 0.99 , 0.   , 0.005],\n",
       "       [0.   , 0.003, 0.987, 0.   , 0.01 ],\n",
       "       [0.   , 0.004, 0.992, 0.   , 0.004],\n",
       "       [0.   , 0.007, 0.992, 0.001, 0.   ],\n",
       "       [0.   , 0.003, 0.928, 0.   , 0.069],\n",
       "       [0.015, 0.   , 0.726, 0.   , 0.258],\n",
       "       [0.   , 0.001, 0.714, 0.   , 0.286],\n",
       "       [0.   , 0.001, 0.706, 0.   , 0.293],\n",
       "       [0.032, 0.   , 0.216, 0.   , 0.752],\n",
       "       [0.022, 0.   , 0.111, 0.   , 0.867],\n",
       "       [0.   , 0.   , 0.106, 0.   , 0.894],\n",
       "       [0.   , 0.   , 0.093, 0.   , 0.907],\n",
       "       [0.002, 0.   , 0.079, 0.   , 0.919],\n",
       "       [0.   , 0.   , 0.12 , 0.   , 0.88 ],\n",
       "       [0.   , 0.   , 0.243, 0.   , 0.756],\n",
       "       [0.019, 0.   , 0.14 , 0.   , 0.84 ],\n",
       "       [0.019, 0.   , 0.138, 0.   , 0.843],\n",
       "       [0.008, 0.   , 0.229, 0.   , 0.763],\n",
       "       [0.177, 0.   , 0.116, 0.   , 0.707],\n",
       "       [0.177, 0.   , 0.23 , 0.   , 0.592],\n",
       "       [0.   , 0.   , 0.946, 0.   , 0.053],\n",
       "       [0.   , 0.002, 0.982, 0.   , 0.016],\n",
       "       [0.   , 0.006, 0.993, 0.   , 0.   ],\n",
       "       [0.   , 0.007, 0.992, 0.001, 0.   ],\n",
       "       [0.004, 0.   , 0.971, 0.   , 0.025],\n",
       "       [0.891, 0.   , 0.068, 0.   , 0.042],\n",
       "       [0.933, 0.   , 0.021, 0.   , 0.046],\n",
       "       [0.941, 0.   , 0.013, 0.   , 0.046],\n",
       "       [0.889, 0.   , 0.066, 0.   , 0.046],\n",
       "       [0.   , 0.001, 0.998, 0.   , 0.001],\n",
       "       [0.041, 0.   , 0.415, 0.   , 0.544],\n",
       "       [0.19 , 0.   , 0.071, 0.   , 0.739],\n",
       "       [0.164, 0.   , 0.025, 0.   , 0.81 ],\n",
       "       [0.008, 0.   , 0.043, 0.   , 0.949],\n",
       "       [0.018, 0.   , 0.009, 0.   , 0.973],\n",
       "       [0.013, 0.   , 0.004, 0.   , 0.983],\n",
       "       [0.007, 0.   , 0.003, 0.   , 0.99 ],\n",
       "       [0.001, 0.   , 0.006, 0.   , 0.993],\n",
       "       [0.   , 0.   , 0.011, 0.   , 0.989],\n",
       "       [0.   , 0.   , 0.008, 0.   , 0.992],\n",
       "       [0.002, 0.   , 0.007, 0.   , 0.991],\n",
       "       [0.003, 0.   , 0.017, 0.   , 0.98 ],\n",
       "       [0.   , 0.   , 0.069, 0.   , 0.931],\n",
       "       [0.   , 0.   , 0.124, 0.   , 0.876],\n",
       "       [0.   , 0.   , 0.118, 0.   , 0.882],\n",
       "       [0.   , 0.   , 0.154, 0.   , 0.846],\n",
       "       [0.029, 0.   , 0.101, 0.   , 0.87 ],\n",
       "       [0.035, 0.   , 0.124, 0.   , 0.842],\n",
       "       [0.03 , 0.   , 0.283, 0.   , 0.687],\n",
       "       [0.   , 0.003, 0.99 , 0.   , 0.007],\n",
       "       [0.   , 0.01 , 0.99 , 0.001, 0.   ],\n",
       "       [0.   , 0.008, 0.992, 0.   , 0.   ],\n",
       "       [0.   , 0.003, 0.985, 0.   , 0.011],\n",
       "       [0.001, 0.001, 0.961, 0.   , 0.037],\n",
       "       [0.   , 0.001, 0.972, 0.   , 0.027],\n",
       "       [0.   , 0.003, 0.994, 0.   , 0.002],\n",
       "       [0.   , 0.006, 0.993, 0.   , 0.001],\n",
       "       [0.   , 0.009, 0.989, 0.   , 0.002],\n",
       "       [0.   , 0.014, 0.967, 0.001, 0.018]])"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.round(model.predict_proba(x1),3)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "pred_prob = np.round(model.predict_proba(x1),3).T\n",
    "hidden_states = model.predict(x1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x720 with 5 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, axs = plt.subplots(model.n_components, sharex=True, sharey=True, figsize=(12,10))\n",
    "colors = cm.Set1(np.linspace(0, 1, model.n_components))\n",
    "\n",
    "for i, (ax, color) in enumerate(zip(axs, colors)):\n",
    "    # Use fancy indexing to plot data in each state.\n",
    "    mask = hidden_states == i\n",
    "    \n",
    "    ax.fill_between(pd.Series(x1.flatten()).index.values[mask],\n",
    "                 np.repeat(80, x1.flatten()[mask].shape[0]),\n",
    "                 x1.flatten()[mask],\n",
    "                 step=\"pre\", \n",
    "                 color = color,\n",
    "                 alpha=0.3)\n",
    "    \n",
    "    ax.plot(pd.Series(x1.flatten()).index.values[mask],\n",
    "            x1.flatten()[mask],\n",
    "            drawstyle=\"steps\", \n",
    "            color = color,\n",
    "            alpha = 0.4)\n",
    "    \n",
    "    ax.plot(pd.Series(x1.flatten()).index.values[mask],\n",
    "                 x1.flatten()[mask],\n",
    "                 \".\", c=color)\n",
    "    ax.set_title(f\"{i}th hidden state\")\n",
    "\n",
    "plt.xlabel('Time')\n",
    "plt.tight_layout()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'Probability of a hidden state')"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1152x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(16,5))\n",
    "\n",
    "for i,s in enumerate(pred_prob):\n",
    "    plt.plot(s, label =str(i))\n",
    "plt.legend(loc = 'upper right')\n",
    "\n",
    "plt.xlabel('Time')\n",
    "plt.ylabel('Probability of a hidden state')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Kết luận\n",
    "\n",
    "Qua thí nghiệm này, chúng ta thấy mô hình Markov ẩn là một phương pháp hữu ích để học thông tin từ dữ liệu chuỗi trong hoàn cảnh ta chưa có một giả thuyết rõ ràng và muốn khảo sát đồng thời cả giá trị quan sát và yếu tố chuyển tiếp, thời gian trong cùng bài toán. Mô hình HMM có thể dùng được cho cả chuỗi giá trị định tính hoặc định lượng, thậm chí có thể mở rộng khảo sát đồng thời nhiều chuỗi song song (thí dụ SpO2 + mạch + huyết áp). \n",
    "\n",
    "Ghi chú: Vì lý do bảo mật, Nhi không chia sẻ dữ liệu, nhưng với nội dung code và cấu trúc dataframe, các bạn có thể dễ dàng tái lập thí nghiệm với dữ liệu mô phỏng hoặc bài toán tương tự."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "hide_input": false,
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  },
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}