README.md

metadata

title: ML Demo - Churn & Recommendations
emoji: 📈
colorFrom: pink
colorTo: yellow
sdk: streamlit
sdk_version: 1.55.0
python_version: '3.10'
app_file: app.py
pinned: true
short_description: ml demo

ML Demo - Customer Churn & Product Recommendations

A comprehensive machine learning demo showcasing:

Features

Customer Churn Prediction

• Exploratory Data Analysis (EDA)
• Multiple ML models: Logistic Regression, Random Forest, XGBoost, Naive Bayes
• SHAP explanations for model interpretability
• Live model updates with streaming data

Product Recommendation System

• Multiple recommendation algorithms: SVD, ALS, SGD (Funk SVD), NMF, Item-Based CF
• Evaluation metrics: Precision@K, Recall@K, Hit Rate, RMSE, R²
• Interactive product recommendations
• Live recommendation updates

Tech Stack

• Streamlit for the web interface
• scikit-learn, XGBoost for ML models
• implicit library for ALS recommendations
• SHAP for model explanations
• Pandas, NumPy for data processing

Dataset

• Telco Customer Churn dataset
• Online Retail dataset (UCI ML Repository)