Weighted Shapley Values and Weighted Confidence Intervals for Multiple Machine Learning Models and Stacked Ensembles

Using a Kaggle dataset, customer personality was analysed on the basis of their spending habits, income, education, and family size. K-Means, XGBoost, and SHAP Analysis were performed.

Análise Avançada de Dados com Causalidade e Aprendizado por Reforço

Análise do Impacto da Padronização de Markdown na Carga Cognitiva e Desempenho de Tarefas

Análise Causal de Intervenções de Ansiedade com Algoritmos de Descoberta Causal

Análise de Intervenção para Ansiedade com Mediação Causal

Análise de Intervenção em Ansiedade com Descoberta Causal

Análise Aprimorada de Intervenção para Ansiedade com LLM Fine-Tuned

This project contains codes and paperwork based on the course CSI5155 at University of Ottawa (delivered by Professor Dr. Herna Viktor).

GLM with sklearn, joblib and SHAP project

🌐 Predicting Customer Churn with Decision Tree, XGBoost & Neural Network Models on the Cell2Cell Dataset

At Infosys Springboard, I worked on a project focused on unsupervised anomaly detection in healthcare providers. I implemented three machine learning algorithms—Isolation Forest, Elliptic Envelope, and One-Class SVM—as well as a deep learning approach using autoencoders. Additionally, I conducted individual SHAP analysis

Binary classification of industrial machine failures using ensemble learning techniques (XGB, LGBM, RF) with SHAP interpretability.

Healthcare AI Assistant Pro is an advanced analytics platform that leverages machine learning and artificial intelligence to predict patient readmission risks. This enterprise-grade solution provides healthcare institutions with data-driven insights to improve patient care, reduce readmission rates, and optimize resource allocation.

Collection of the assignments for Data Science Engineering Methods on National Stock Exchange Dataset and TMNIST dataset

Reproducible ML pipeline evaluating temporal leakage in Expected Pass Turnovers (xPT) models for football analytics. Compares 4 algorithms (mixed-effects logistic, penalised logistic, random forest, XGBoost) across leakage-inclusive and leakage-corrected feature sets. Supporting code for manuscript under review.

This repository contains the Python implementation for the article "Integrating Transformers and Gaussian Mixture Models for Parkinson's Detection." The code combines GMM, Transformers, and SHAP analysis for accurate and interpretable voice-based diagnosis.

Predict drug sensitivity (IC50) using a Dual-Branch Deep Learning model to solve the "Cold Start" problem in precision medicine. Features: GDSC dataset, SHAP explainability, and Optuna hyperparameter tuning.

A comprehensive machine learning pipeline for cardiovascular disease prediction using deep neural networks with explainable AI capabilities.