Hyperparameters-Optimization

Steps for hyperparameter tuning using Random, Grid CV and Hyperopt

A machine learning pipeline for classifying cybersecurity incidents as True Positive(TP), Benign Positive(BP), or False Positive(FP) using the Microsoft GUIDE dataset. Features advanced preprocessing, XGBoost optimization, SMOTE, SHAP analysis, and deployment-ready models. Tools: Python, scikit-learn, XGBoost, LightGBM, SHAP and imbalanced-learn

Selected Paper from the AI-CyberSec 2021 Workshop in the 41st SGAI International Conference on Artificial Intelligence (MDPI Journal Electronics)

Inter class kaggle Hackathon.

Project made for Optimisation and Deep Learning course.

Bank Fraud Detection with Imbalanced Data: Applying Oversampling and Hyperparameter Optimization

Implementing k-fold random search cross validation from scratch for a KNN classifier.

GUI for hyperparameter optimizer and plotting experiment results

Advanced Machine Learning

Using a synthetic dataset from Kaggle, generated with Python's Faker library to mimic real Twitter data, we train several classical machine learning models (ie. classical classification algorithms, as well as ensemble methods)to identify bots from real users.

A comprehensive set of programs demonstrating machine learning techniques have been made.

The project is aimed at predicting the rate of penetration for natural gas and oil mines. The project was made as a part of the Petrocoder National Challenge 2021 and achieved a national rank of 4 in the competition.

Ensemble Learning with Pima_Indian dataset