Global partial orders from sequential data
Heikki Mannila, Christopher Meek · 2000
Sequences of events arise in many applications, such a s w eb browsing, e-commerce, and monitoring of processes.An importan t problem in mining sets of sequences of ev ents is to get an o verview of the ordering relationships in the data.W e presen t a method for nding partial orders that describe the ordering relationships between the events in a collection of sequences.The method is based on viewing a partial order as a generative model for a set of sequences, and applying mixture modeling techniques to obtain a descriptive s e t o f partial orders.Runtimes for our algorithm scale linearly in the number of sequences and polynomially in the number of dierent e v ent t ypes.Thus, the methods scales to handle large data sets and can be used for reasonable numbers of dierent t ypes of events.We illustrate our technique by applying it to studen tenrollment data and web browsing data.