Mining Frequent Patterns from Spatio- Temporal Data Sets: a Survey

Space and time are implicit in every activity of life. Every real-world object has its past, present, future and hence is intrinsically tied up with location and time. Storing spatio-temporal attributes in the databases along with the thematic attributes enriches the data and the inherent knowledge stored in the database. Spatio-temporal databases provide description of real-world phenomenon in a spatio-temporal framework and helps to build information systems and services that are more comprehensive and intelligent. This paper highlights the importance and applications of spatio-temporal pattern mining and provides a brief survey of key mining techniques for discovering three types of spatio-temporal patterns – sequential patterns, cooccurrence patterns and cascaded patterns specifically from event data sets and trajectory data sets.