Skip to content

Latest commit

 

History

History
60 lines (35 loc) · 1.84 KB

PROPOSAL.md

File metadata and controls

60 lines (35 loc) · 1.84 KB
Raw

Classification with class imbalance problem: A review

Team Details

Course : [CS7.403] Statistical Methods in AI
Instructor : Dr. Anoop Namboodiri
Team Name : Skynet
Participants :

  • Megha Bose
  • Samyak Jain
  • Nirmal Manoj
  • Varul Srivastava

Problem Statement

Goals and Approach

Data-Driven Approach

SMOTE : Synthetic Minority Over-sampling TEchnique

This paper shows that a combination of our method of over-sampling the minority (abnormal) class and under-sampling the majority (normal) class can achieve better classifier performance (in ROC space) than only under-sampling the majority class.

This paper also shows that a combination of our method of over-sampling the minority class and under-sampling the majority class can achieve better classifier performance (in ROC space) than varying the loss ratios in Ripper or class priors in Naive Bayes.

Fig.1 SMOTE Approach

DATASET : We plan to use the Pima Indians Diabetes Database, which contains 2 classes and 768 samples. The reason for choosing this was, that it's a medical database and thus, is a realistic scenario where such class imbalances occour. Number of positive class samples is only 268.
Link-to-Dataset

Algorithm Driven Approach

Expected Deliverables

Milestones and Timeline

Timeline Milestone
November 2 Project Allocation
November 7 Project Proposal Finalized
November 14 Implementation of 4 of 8 Approaches
November 21 Implementation complete
November 22 Project report complete
November 24 Project Completion

References