GitHub

An end-to-end banking marketing prediction pipeline using SQL for deduplication and feature engineering, and Python (Gradient Boosting/LightGBM) to handle imbalanced data and ...

The dataset used in this project is the Bank Marketing (UCI) dataset, sourced from Kaggle. Dataset Link: https://www.kaggle.com/datasets/henriqueyamahata/bank ...
GitHub

GitHub - epiasoo/Road_Accident_Prediction_Machine_Learning: Predictive modelling of UK road accident severity using LightGBM with SMOTENC for class imbalance handling.

This project develops an interpretable and scalable machine learning framework to predict road accident severity (Fatal, Serious, Slight) using real-world UK traffic collision data. The model deploys ...
IEEE

SHAP-LM: Optimized Music Genre Classification via SHAP-Guided Feature Selection and Lightweight Modeling

Abstract: Music genre classification (MGC) is a fundamental task in music information retrieval, yet existing approaches often suffer from high computational cost, limited generalization, and lack of ...
IEEE

Enhanced Accident Severity Prediction Using LightGBM-CNN Fusion Framework

Abstract: This research presents a novel hybrid approach combining Convolutional Neural Networks (CNN) with Light Gradient Boosting Machine (LightGBM) for traffic accident prediction. The proposed ...