Machine Learning Projects

This repository contains a collection of machine learning projects spanning multiple domains, including business applications, power systems & cybersecurity, and sensor-based activity recognition. Each notebook demonstrates different ML techniques, ranging from classical models to explainable AI and adversarial robustness.

---

Repository Structure

machine-learning/ │── README.md │── requirements.txt │ ├── business_ml/ # Business-focused ML projects │ ├── credit_card_fraud_detection.ipynb │ ├── customer_segmentation_kmeans.ipynb │ ├── netflix_stock_price_prediction_arima.ipynb │ ├── customer_churn_prediction_lime.ipynb │ ├── privacy_preserving_churn_prediction_dp.ipynb │ ├── linear_regression_prediction.ipynb │ ├── image_classification_svm.ipynb │ ├── power_systems_cybersecurity/ # ML for energy systems & cybersecurity │ ├── intrusion_detection_systems_ids.ipynb │ ├── iot_device_anomaly_detection.ipynb │ ├── smart_grid_predictive_maintenance.ipynb │ ├── adversarial_robustness_ids.ipynb │ ├── powergrid_ids_ml.ipynb │ ├── tsne_pca_energy_consumption.ipynb │ ├── sensors_wearables/ # Activity recognition with sensor data │ ├── activity_recognition_smartphone_sensors.ipynb

---

Projects Overview

🔹 Business ML

• Credit Card Fraud Detection – ML for fraud detection in financial transactions.
• Customer Segmentation (K-means) – Cluster analysis for targeted marketing.
• Netflix Stock Price Prediction (ARIMA) – Time series forecasting.
• Customer Churn Prediction (with LIME) – Explainable churn modeling.
• Privacy-Preserving Churn Prediction (Differential Privacy) – Privacy-aware ML.
• Linear Regression Prediction – Classical regression modeling.
• Image Classification using SVM – Baseline ML for image data.

🔹 Power Systems & Cybersecurity

• Intrusion Detection Systems (IDS) – Security in smart grids.
• IoT Device Anomaly Detection – Detecting abnormal device behavior.
• Smart Grid Predictive Maintenance – Sensor-based predictive analytics.
• Adversarial Robustness in IDS – Testing ML security.
• PowerGrid IDS (ML-based) – Intrusion detection in energy networks.
• TSNE & PCA on Energy Consumption – Dimensionality reduction.

🔹 Sensors & Wearables

• Activity Recognition (Smartphone Sensors) – Human activity classification using wearable data.

---

Requirements

To install dependencies, run:

pip install -r requirements.txt