Effect-Coding-and-Regression-Analysis

https://benthamchang.github.io/Effect-Coding-and-Regression-Analysis/Effect%20Coding%20and%20Regression%20Analysis.html

Goal

Classify candidates into non-stroke and stroke patients based on the inhibition between Tibialis Anterior and Soleus muscles under limited data points.

Analysis

1. Exploratory Data Analysis
2. Built a second-order WLS polynomial regression model with effect coding technique applied to capture the relationship between the powers of Tibialis Anterior and Soleus muscles.
3. Classified candidates into non-stroke and stroke patients according to the coefficients' values of above model.

Conclusion

Obtained 100% accuracy by succeeding detecting 3 stroke patients out of 17 candidates.