Tyred / TimeSeries_OCC-PUL

One-Class Classification and Positive-Unlabeled Learning algorithms for Time Series

One-Class Classification (OCC) and Positive-Unlabeled Learning (PUL) for Time Series Data

Undergraduate Research supported by FAPESP Scientific Initiation Scholarship.

This research aims to study time series classification techniques' applicability in the one-class classification and positive-unlabeled learning tasks. The goal of this repository is to store both the source codes developed and the results obtained in this research. More info about this research can be found here.

To test and evaluate the algorithms, we use the datasets contained in the UCR Time Series Classification Archive.