The subject of this repository was to perform binary classification based on patient's collected features (age, cholesterol level, fasting blood sugar, thallium stress test results, etc.). We implemented different machine learning models which can categorize these labels. We used supervised learning with the scikit-learn library. These algorithms were developed to classify target labels:
- logistic regression,
- k-nearest neighbors,
- support vector machines,
- random forest ensemble,
- Gaussian Naïve Bayes,
- gradient boosting ensemble,
- stochastic gradient descent,
- multi-layer perceptron.
We achieved such results while evaluating the models. The top classifier with highest performance was the k-nearest neighbors. We prooved that forecasting heart disease events with this dataset is possible.