HexDoc 🩺

HexDoc is an AI-powered chatbot designed to offer personalized stress relief exercises and mental wellness resources. Utilizing basic Natural Language Processing (NLP), HexDoc interprets user inputs related to mood and stress levels to provide tailored recommendations. The chatbot ensures ease of interaction and engages users in a calming, supportive conversation, helping them manage stress effectively.

Functionalities 📃

Natural Language Processing (NLP):

• Utilized basic NLP techniques to understand user inputs and respond accordingly. • The chatbot interprets simple prompts related to stress levels, emotions, or mood.

Motivational and Supportive Content:

• Motivational quotes, short wellness tips, and access to mental health resources as needed.

Ease of Interaction:

• Chatbot is user-friendly and responsive, making it easy for users to engage with the chatbot during stressful moments.

Conversational Flow:

• Chatbot engages users in a calming, supportive conversation, offering a range of personalized options depending on the input received.

Depression Test and Analysis

• Evaluates depression symptoms using a structured questionnaire, providing a score to gauge the severity of symptoms.

Anxiety Test and Analysis

• Assesses anxiety levels through targeted questions, calculating a score to determine the intensity of anxiety symptoms.Delivers tailored feedback and coping strategies based on the results, with links to additional resources and professional support.

Project Installation 🚀

Initialization

Create project directory

mkdir hexdoc

Move to project directory

cd hexdoc

Installation

Clone the repository

git clone https://https://github.com/lohithgsk/chatbot-rag.git

Installing dependencies

pip install -r requirements.txt

Usage

Start the flask application on your terminal

flask run

The application would be up at https://localhost:5000

About the model 🤖

This is a feedforward neural network built using TensorFlow's Keras API for classifying user inputs into predefined categories or intents. Here's a brief description:

• Input Layer: The model takes in a vector of length equal to the number of unique words in the training data (bag-of-words representation).
• First Hidden Layer: The first layer has 128 neurons with ReLU activation, introducing non-linearity to capture complex patterns in the data. A dropout layer follows, randomly dropping 50% of the neurons during training to prevent overfitting.
• Second Hidden Layer: This layer has 64 neurons with ReLU activation, followed by another dropout layer, further refining the model's ability to generalize.
• Output Layer: The output layer has as many neurons as there are classes (intents) in the dataset, with a softmax activation function to output a probability distribution over all possible classes.
• Compilation: The model is compiled with categorical cross-entropy as the loss function (suitable for multi-class classification) and the Adam optimizer with a learning rate of 0.001, optimizing for accuracy.

This architecture is designed to classify input text into specific categories, such as different types of user intents, by learning from labeled training data.

Project Screenshots 📸

Home page of the application

Chatting interface

Mental health test - home page

Depression Test

Depression Test Result

Anxiety Test

Anxiety Test Result

Feedback 💬

If you have any feedback you can reach out to us.