search

yzhao062 / pyod

A Python Library for Outlier and Anomaly Detection, Integrating Classical and Deep Learning Techniques
http://pyod.readthedocs.io
BSD 2-Clause "Simplified" License
8.56k stars 1.37k forks source link

Optimal model for multivariate categorical time series data. #202

Open michaelTucci opened 4 years ago

michaelTucci commented 4 years ago

I have a pretty complicated dataset that I'd like to run supervised or unsupervised OD on, but I'm at a loss on where to start in terms of models.

Basically, I have a bunch of categorical multivariate time series sets where each time series is associated with a distinct individual person. The sets are not uniform in length nor are the timestamps of each vector at regular intervals. My final product would ideally be a model that runs through a new person's time series and marks outlying vectors.

The data I have can be easily labeled, so I'd like to try both supervised and unsupervised learning. Any thoughts?

NeighborhoodCoding commented 4 years ago

good idea, how to classify time series data? ^.^?