View Proposal
-
Proposer
-
Arulmurugan Ramu
-
Title
-
Assessing Deep and Classical Anomaly Detection Techniques for Tabular Data
-
Goal
-
-
Description
- This project examines the use of deep learning and classical machine learning models for detecting anomalies in tabular data. By comparing several detectors under a controlled evaluation protocol and measuring how far each degrades when the assumed contamination rate is wrong or may be methodological critique .
- Resources
-
-
Background
-
[1] Han, S., et al. "ADBench: Anomaly Detection Benchmark." NeurIPS 2022 Datasets and Benchmarks. https://github.com/Minqi824/ADBench
-
Url
-
-
Difficulty Level
-
Easy
-
Ethical Approval
-
None
-
Number Of Students
-
1
-
Supervisor
-
Arulmurugan Ramu
-
Keywords
-
anomaly detection, tabular data, deep learning, contamination rate, threshold selection,evaluation methodology
-
Degrees
-
Bachelor of Science in Computational Sciences and Software Engineering