Spotify Machine Learning Project

📌 Project Summary

This notebook explores a Spotify dataset to predict listening behavior based on track features.

1. Data Preparation

• The dataset was cleaned, preprocessed, and standardized using StandardScaler to improve model performance.

2. Model Training & Evaluation

• Multiple machine learning models were trained and compared, including Logistic Regression, Decision Trees, Random Forest, KNN, XGBoost, and others.
• Accuracy and cross-validation were used to evaluate performance.

3. Model Selection

• Random Forest consistently gave the highest accuracy and balanced performance compared to other models.

4. Hyperparameter Optimization

• Optuna was used to fine-tune Random Forest parameters (n_estimators and max_depth).
• Cross-validation ensured the model was not overfitting.

5. Final Results

• Best Model: Random Forest Classifier
• Best Accuracy: 0.9892
• Best Parameters: n_estimators = 91, max_depth = 27

---

✅ The project successfully identified Random Forest as the optimal model and optimized it for strong predictive performance.