LoTTA: Low-rank Test-time Adaptation for Unsupervised Tabular Anomaly Detection

The code of paper "LoTTA: Low-rank Test-time Adaptation for Unsupervised Tabular Anomaly Detection"，Information Processing & Management. IPM 2026

This repository is not yet completed, so please check this as a reference only.

Prepare dataset

1. When using your own data, move the dataset into ./Data.
2. Add the dataset name to ./Dataset/DataLoader.py based on the format of your dataset.
3. You can download tabular datasets from ODDS and ADBench for testing.