A Proposal of Spatio-Temporal Pattern Queries

This paper provides a comprehensive analysis of algorithms for spatio-temporal pattern mining, which helps to select appropriate algorithms for a given Moving Objects Database. We can distinguish basic and extended comparing functions. The basic ones correspond to research fields related to pattern mining. The extended comparing functions extend the analysis to additional aspects. All these functions are presented in the context of pattern queries for Moving Objects Databases. In addition, comparing functions can be combined into more complex ones and used in frequent pattern mining algorithms in order to obtain various summaries of all analyzed sequences. We also present two algorithms, called VGES (Vertical GSP for ExactSearch) and SES (Spade for ExactSearch), for continuous sequential pattern mining. In the paper different methods of searching for frequent patterns are also compared.