Dataset for this project is taken from UCI ML Repository: https://archive-beta.ics.uci.edu/dataset/519/heart+failure+clinical+records
This project is based on a binary classification task to identify if the patient is a high risk patient. I compared different classification algorithms on their Accuracy, Precision, Recall and F1 Score. I started by analysing the data, and then trained and tested the models on Logistic Regression, K-Nearest Neighbors, Support Vector Classifier, Decision Tree Classifier, and Random Forest Classifier. I used the Scikit-learn library to implement and evaluate each algorithm.
After testing all the algorithms, the best metrics achieved were Accuracy - 0.8, Precision - 0.88, Recall - 0.6 and F1 Score - 0.71.
After working on this project, I learned the importance of comparing and evaluating different algorithms across important metrics to get a robust and reliable model.