Computing Reviews

MixedTrails:Bayesian hypothesis comparison on heterogeneous sequential data
Becker M., Lemmerich F., Singer P., Strohmaier M., Hotho A. Data Mining and Knowledge Discovery31(5):1359-1390,2017.Type:Article
Date Reviewed: 12/07/17

This well-written paper includes adequate definitions to enable a layperson to understand the principles (generative processes of heterogeneous sequence data of human movement in a city) examined in its simulated experimental study. It uses appropriate examples, especially Figure 1 (the soccer example), to illustrate the model.

The comparisons between different sequences of events by modeling them as Markov chains and obtaining probabilities from a Bayesian model are contributions of the paper. By doing this, the authors demonstrate the differences between higher-level strategies (for example, offensive) that are implemented as different sequences. The higher-order strategies also require knowledge of the time between events in the sequence, not solely the event sequences. It is understood that the contribution is based on experimental results. Overall, this is a top paper with a valid scientific argument to share.

Reviewer:  Tony Sahama Review #: CR145698 (1802-0095)

Reproduction in whole or in part without permission is prohibited.   Copyright 2024 ComputingReviews.com™
Terms of Use
| Privacy Policy