This repository comprises the results of my thesis project:
Outlier detection in multivariate time series: exploiting reconstructions from random projections.
This work has been produced under supervision of David Tax, at the Pattern Recognition Laboratory of Delft University of Technology.
The contents of this repository are as follows:
- The data used for the analysis of the two proposed Random Projection methods, and real-world experiments.
- The code that has been used to generate the results following the setup as explained in the thesis. The code comes in the form of four functions:
- RP_RECON: the original default RP method.
- RP_RECON_SCALED: the default RP method with back-scaling.
- RP_RECON_DELTA: the deltaRP method.
- RP_RECON_DELTA_OFFLINE: the offline version of the deltaRP method.
- The files used to generate the thesis document.
- The digital final version of the thesis.
I officially started my thesis around 20 November 2017 and will defend my work on 20 July 2018 at 9:30 AM.