Implementation of the Geometric SMOTE over-sampling algorithm.

Implementation of novel oversampling algorithms.

Malicious URL detector built with deep exploration on feature engineering.

We leverage machine learning and data analysis to address real-world challenges in the copper industry. Our documentation encompasses data preprocessing, feature engineering, classification, regression, and model selection. Explore how we've enhanced predictive capabilities to optimize manufacturing solutions.

Python package for tackling multi-class imbalance problems. http://www.cs.put.poznan.pl/mlango/publications/multiimbalance/

Scoring model for financial company - all files

Synthetic Minority Over-Sampling Technique for Regression

Routines to perform cross-validation and nested cross-validation using data transformations

Continuing with telemarketing model to predict campaign subscriptions in a portuguese bank institution. For this project I have evaluated the performance of four resampling techniques and selected the best one to implement the logistic model.

A telemarketing model to predict campaign subscriptions in a portuguese bank institution.

Fraud Machine Learning Pipelining for experimenting with SMOTE

demonstrate different models such as Variational Autoencoders and GANs in a variety of datasets, including tabular, text and image data, including the generation of synthetic data for comparison of their effectiveness in all models for each kind of dataset

Used CDC dataset for heart attack detection in patients. Balanced the dataset using SMOTE and Borderline SMOTE and used feature selection and machine learning to create different models and compared them based on metrics such as F-1 score, ROC AUC, MCC, and accuracy.

This repository contains the code of our published work in IEEE JBHI. Our main objective was to demonstrate the feasibility of the use of synthetic data to effectively train Machine Learning algorithms, prooving that it benefits classification performance most of the times.

Data warehouse and analytics project to predict bike theft prediction from TPS data

Scripts developed for the paper "Understanding when SMOTE works", developed in the "Knowledge Extraction and Machine Learning" (ECAC) class.

The computing scripts associated with our paper entitled "Oversampling Highly Imbalanced Indoor Positioning Data using Deep Generative Models".

The experimental codes using PyTorch from the paper that was submitted to GECCO 2020.

A minority oversampling method for imbalance data set