[NeurIPS 2022] Source code for our paper "Escaping Saddle Points for Effective Generalization on Class-Imbalanced Data"

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

SMAI Project of Team Skynet: Analysing methods to tackle class imbalance for classification problems

This repository contains the complete project for Emotion Detection using Convolutional Neural Networks (CNNs) and the FER-2013 dataset. The project focuses on addressing class imbalance, applying data augmentation techniques, and leveraging advanced architectures such as VGG16 and ResNet50v2 to improve the robustness and accuracy.

model to predict whether a customer would be interested in vehicle insurance

Performed feature engineering, cross-validation (5 fold) on baseline and cost-sensitive (accounting for class imbalance) Decision trees and Logistic Regression models and compared performance. Used appropriate performance metrics i.e., AUC ROC, Average Precision and Balanced Accuracy. Outperformed baseline model.

Build a CNN based model which can accurately detect melanoma

RCSMOTE: Range-Controlled Synthetic Minority Over-sampling Technique for handling the class imbalance problem

Classification of movies as 'Fresh', 'Rotten', 'Certified-Fresh' using categorical predictors as well as review sentiment. Performed feature encoding and used Decision Tree, Random Forest Classifiers. Tackled class imbalance issues by assigning weights to classes. Used tokenization to generate word vectors for reviews to predict movie status.

This code is a PyTorch implementation of ClassAwareLoss proposed in the "Class-aware fish species recognition using deep learning for an imbalanced dataset" paper. https://www.mdpi.com/1424-8220/22/21/8268

Develop Machine Learning model to predict customer loan defaults, enhancing lending risk assessment. Real-world relevance tackling financial uncertainty. #The Analytics Olympiad 2023

Code for "Generative Oversampling for Imbalanced Data via Majority-Guided VAE", AISTATS2023.