Portfolio Details
Customer Insights and Recommendation System
Developed a Customer Insights and Recommendation System using Decision Tree, Logistic Regression, and Random Forest models, achieving 82% accuracy in predicting customer conversions. Extracted key insights from transaction data to improve campaign engagement by 15% and built a recommendation pipeline that boosted retention by 10%.
Built a unified Customer Insights system using ML models, OCR, and sentiment analysis to predict conversions and generate personalized product recommendations.
Managing mixed e-commerce data (behavioral, image, and text) while ensuring accurate conversion prediction and real-time recommendations.
Processed multi-modal data, trained ML models, applied OCR and sentiment analysis, and deployed a Streamlit app for live predictions and recommendations.
Key Features
- Conversion Prediction
- precision/recall/accuracy/F1
- Image OCR & Processing:
- Recommendation Engine:
- Text Sentiment Analysis
- EDA