Predicting customer retention in an e-commerce platform using classification models. Includes data preprocessing, feature engineering, and model evaluation (Logistic Regression, SVM, Random Forest, KNN, Decision Tree). Best model achieves 83% accuracy and perfect recall. Ideal for business use.
This project converts a Jupyter-based machine learning model into a modular, cloud-ready data engineering pipeline using Python, AWS S3, and PostgreSQL. It enables automated data ingestion, transformation, and loading of credit card transaction data for downstream analysis or modeling.
Spam Email Classification using LoRA Fine-tuned Transformers: High-performance spam email classification using LoRA-adapted transformer models (ELECTRA, RoBERTa). Achieves 99.4%+ accuracy with parameter-efficient fine-tuning on 83K+ emails.
A robust end-to-end machine learning pipeline for credit card fraud detection using Python, scikit-learn, and Streamlit. Includes data preprocessing, feature selection, model training & evaluation, saving the best model, and an interactive Streamlit app for predictions.
To associate your repository with the imbalanced-dataset-handling topic, visit your repo's landing page and select "manage topics."