📊 Dataset

Source: (Heart Disease Dataset)[https://www.kaggle.com/datasets/redwankarimsony/heart-disease-data]

Description: The dataset includes various clinical features such as age, cholesterol levels, blood pressure, and more, aimed at determining the presence of heart disease.

🚀 Features

Data Preprocessing:

Handling missing values using different methods like iterative imputer or filling by constant value, encoding categorical variables, and feature scaling.

Feature Selection:

Application of Recursive Feature Elimination (RFE) and Chi-Square tests to identify significant predictors. I also plotted each feature importance.

Dimensionality Reduction:

Used PCA to preserve the most variance for Unsupervised Machine learning models.

Model Training:

Supervised:

Implementation of multiple classifiers including Logistic Regression, Random Forest, Decision Tree, and Support Vector Machine (SVM).

UnSupervised:

Kmeans and Hierarchical Clustering to group similar objects.

Hyperparameter Tuning:

Used GridSearch and RandomizedSearch to get the best paramters for models.

• Logistic Regression: 82% Accuracy
• Decision Tree: 80% Accuracy
• Random Forest: 84% Accuracy
• SVM: 84% Accuracy

Model Evaluation:

Utilization of metrics like ROC curves and AUC scores to assess model performance.

Model Persistence:

Saving trained models using joblib for future inference.

🛠️ Installation

1. Clone the repository:

git clone https://github.com/yahia997/SprintUp-AI-ML-project.git
cd SprintUp-AI-ML-project

2. Install dependencies:

pip install -r requirements.txt