HypGraphsHypGraphs

This extension implements the HypGraphs modeling and analysis approach proposed by Atzmueller et al. (2016). It calculates the evidence score that a hypothesis transition matrix is similar to a given data matrix. For interference calculation the HypTrails algorithm (Singer et al. 2015) is used, by the implementation of Florian Lemmerich (https://bitbucket.org/florian_lemmerich/hyptrails4j).

Description

This exctension implements HypGraphs modeling and analysis approach proposed by Atzmueller et al. (2016) to compare two transition matrices. It can be used to discern if two patterns of sequences have a similar structure or not.
With that it can be used to detect anomalies on sequential data.

The extension includes two operators:

Attribute to Transition Matrix to transform an attribute of events into a transition matrix

HypGraphs to calculate the evidence score that two transition matrices are similar

 

Literature
Atzmueller et al. (2016): Atzmueller, Martin; Schmidt, Andreas; Kloepper, Benjamin and Arnu, David (2016) HypGraphs: An Approach for Modeling and Comparing Graph-Based and Sequential Hypotheses. Proc. ECML-PKDD Workshop on New Frontiers in Mining Complex Patterns (NFMCP), Riva del Garda, Italy, 2016

Singer et al. (2015): Singer, Phillip; Helic, Denis; Hotho Andreas and Strohmaier Markus (2015) HypTrails: A Bayesian Approach for Comparing Hypotheses About Human Trails on the Web, Proc. International World Wide Web Conference, Florence, Italy, 2015


Product Details

Version
File size
Downloads 0 (0 Today)0 downloads
Vendor RapidMiner Labs
Category Machine Learning
Released
Last Update
(Changes)
License
Product web site
Rating 0.0 stars(0)