TSLTO

TSLTO (Tucker decomposition-based Sparse Low-Rank High-order Tensor Optimization Model) is a model for tensor imputation and anomaly diagnosis. Specific model and evaluations can be found in the paper Traffic Flow Data Completion and Anomaly Diagnosis via Sparse and Low-Rank Tensor Optimization.

Quik Run Guide

:star: Don't forget to load Tensor_Toolbox and Tensorlab first!

About Synthetic Data

If you are interested in generating our synthetic data, you can run TSLTO/synthetic_dataset.m .

About Real-World Data

Our model can also handle imputation and anomaly diagnosis tasks in real-world data (we only use Guangzhou).
You first need to run

load(guangzhou.mat);

then run TSLTO/GUANGZHOU.

Contact Us

Junxi Man 22271014@bjtu.edu.cn
Yumin Lin 21261047@bjtu.edu.cn
For more information, feel free to ask!