This is a very Important part of Data Science Case Study because Detecting Frauds and Analyzing their Behaviours and finding reasons behind them is one of the prime responsibilities of a Data Scientist. This is the Branch which comes under Anamoly Detection.

It is a Problem Which I got During the ZS Data Science Challenge From Interview Bit Hiring Challenge Where I secured a 40th Rank out of 10,000 Students across India. It is a Dataset which requires Intensive Cleaning and Processing. Here I have Performed Classification Using Random Forest Classifier and Used Hyper Tuning of the Parameters to achi…

Exploring categorical features with various encodings and models

A python machine learning library for advanced feature extraction and interpretation.

It contains the code and data for M5 Forecasting - Accuracy competition on Kaggle.

A set of tools for machine learning (for the current day, there are active learning utilities and implementations of some stacking-based techniques).

ML system predicting Indian domestic airfares (R2 0.9165, 94.1% accuracy) and recommending cost-effective bookings, with validated ranking reliability

Qualcomm-Institute 14th Winter project @ UC San Diego

A comprehensive Exploratory Data Analysis (EDA) on holiday sales data using Python to uncover demographic purchasing patterns and actionable business insights.

Creating a sophisticated web application for transaction analysis, incorporating ML, Bootstrap, Dash, and Plotly. Users can seamlessly upload credit card CSV files, exploring transactions interactively in both tabular and dashboard report formats.

Production-ready ML pipeline demonstrating feature engineering, hyperparameter tuning, stacking ensembles and pseudo-label learning.

This repository is a comprehensive guide to different Encoding techniques in Machine Learning, explaining when to use each method and best practices. You'll find practical examples, ready-to-use code, and comparisons between various techniques like Label Encoding, One-Hot Encoding, Target Encoding, and more!

End-to-end consumer credit default risk model using LightGBM ; feature engineering, target encoding with PSI drift checks, SHAP explainability, probability calibration, and decile-lift risk segmentation.

Home Credit Risk Prediction (AUC 0.794 / Private LB Top 15%)

Machine learning pipeline for predicting trucking load rates from lane, equipment, market, and temporal features. Includes preprocessing, feature engineering, model training (Ridge, Random Forest, Gradient Boosting, HistGradientBoosting), prediction generation, and submission validation with scoring.

Home Credit Risk Prediction (AUC 0.794 / Private LB Top 15%)

HackerEarth Machine Learning challenge: Of Genomes And Genetics

A machine learning model to accurately predict house prices based on various features such as quality, size, and location, utilizing Random Forest and XGBoost algorithms (Python) (Done alongside Kaggle ML courses)

Hybrid stacking system: [MLP, VAE, LTSM XGBoost, LightGBM, CatBoost, RandomForest] -> LogReg