A non-time series approach to vehicle related time series problems

Wells, J.R., Ting, K.M. and Naiwala, C.P.

This paper shows that some time series problems can be better served as non-time series problems. We used two unsupervised learning anomaly detectors to analyse a vehicle related time series problem and showed that non-time series treatment produced a better outcome than a time series treatment. We also present the benefits of using unsupervised methods over semi-supervised or supervised learning methods, and rule-based methods.

Cite as: Wells, J.R., Ting, K.M. and Naiwala, C.P. (2012). A non-time series approach to vehicle related time series problems. In Proc. Data Mining and Analytics 2012 (AusDM 2012) Sydney, Australia. CRPIT, 134. Zhao, Y., Li, J. , Kennedy, P.J. and Christen, P. Eds., ACS. 61 - 70

(from crpit.com) (local if available)