Small changes

Author: cetagostini-wise

causalpy/pymc_experiments.py

@@ -347,7 +347,8 @@ def _power_estimation(self, alpha: float = 0.05, correction: bool = False) -> Di
         Estimate the statistical power of an intervention based on cumulative and mean results.
         This function calculates posterior estimates, systematic differences, credible intervals, and
         minimum detectable effects (MDE) for both cumulative and mean measures. It can apply corrections to
-        account for systematic differences in the data.
+        account for systematic differences in the data if the mean pre-intervention is consider far from the
+        real mean.
 
         Parameters
         ----------

docs/source/notebooks/sc_power_analysis.ipynb

@@ -24,7 +24,6 @@
     "Our proposed power (sensitivity) analysis method is designed to be universally applicable across different regression models in `causalpy` such as **synthetic controls** and **interrupted time series**. By assessing the null model's posterior, we can validate that our regression does not inadvertently capture effects during a control period. This validation process not only aids in determining the required effect size for significance but also helps in evaluating the natural bias of the model, thus ensuring more reliable and accurate experimental planning and analysis.\n",
     "\n",
     "\n",
-    "\n",
     "### Similarities and Differences to Frequentist Methods\n",
     "\n",
     "Both Bayesian and frequentist methods aim to provide insights into the effectiveness of interventions or treatments. Our power analysis method in the Bayesian context differs from traditional frequentist approaches by focusing on the probability distributions of outcomes instead of point estimates and p-values. Frequentist methods rely on p-values and confidence intervals to reject or fail to reject a null hypothesis. In contrast, Bayesian approaches use the posterior distribution to estimate the probability of various outcomes, providing more nuanced insights into the expected effects and giving the observer the possibility to determine their significance based on the risk of false positives."
@@ -7068,7 +7067,9 @@
     "\n",
     "## Update the model\n",
     "\n",
-    "Following the shortcomings of our previous model, we have incorporated new regressors to improve its efficacy. These additional features are instrumental in assessing whether the updated model achieves greater accuracy and precision in mirroring reality. This enhancement is crucial for more accurately estimating experimental outcomes."
+    "Following the shortcomings of our previous model, we have incorporated new regressors to improve its efficacy. These additional features are instrumental in assessing whether the updated model achieves greater accuracy and precision in mirroring reality. This enhancement is crucial for more accurately estimating experimental outcomes.\n",
+    "\n",
+    "During this same process we can iterate to determine the model parameters and configurations that give us a better representation of reality during said period."
    ]
   },
   {