Project Overview

This project aimed to predict the likelihood of an individual experiencing a stroke by utilizing various classification models and analyzing ten independent variables. The research question focused on assessing how effectively these variables could predict whether an individual had a stroke or not.

To achieve this, several classification tests, SVM models, and decision trees were employed using R:

After thorough analysis, the LDA model emerged as the most effective in predicting stroke likelihood with the highest AUC value. LDA likely demonstrated superior performance due to its ability to find the linear combinations of features that best separate the classes, making it well-suited for this classification task.

This project underscores the importance of employing a variety of classification techniques when tackling predictive modeling tasks. By systematically evaluating multiple models, researchers can identify the most suitable approach for a given dataset and research question. Additionally, understanding the strengths and weaknesses of each model aids in interpreting and contextualizing the results, ultimately enhancing the reliability and applicability of the predictive model.