Abstract | ||
---|---|---|
Sequence learning consist of finding sequence patterns following an inclusion relation regarding the ordering of events. This is a process that requires hard computational time, in particular when the lengths of the sequences are large. Therefore most of the research is focused on obtaining fast, more efficient algorithms. However, the succession of events has also some properties, as for example, the temporal distance among two events or gap, that should be taken into account. Such gaps could be the key in complex event processing systems to, for example, firing rules regarding fraud detection. This paper tackles the problem of sequence learning when information about the gap between two consecutive events is taken into account. We follow a constraint programming approach to formulate and solve the problem. The experiments are carried out with data from the Santander Cycle bike hiring system to find pattern behaviours of customer station itineraries, which are used to predict the next station a customer will visit. |
Year | DOI | Venue |
---|---|---|
2015 | 10.3233/978-1-61499-578-4-263 | Frontiers in Artificial Intelligence and Applications |
Keywords | DocType | Volume |
Data Mining,Sequence learning,Bike systems | Conference | 277 |
ISSN | Citations | PageRank |
0922-6389 | 1 | 0.37 |
References | Authors | |
0 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Pablo Gay | 1 | 59 | 7.76 |
Beatriz López | 2 | 319 | 42.30 |
Joaquím Meléndez | 3 | 111 | 15.29 |