A multi-phase approach for classifying multi-dimensional sequence data

We present a new methodology for sequential classification, which employs sequential pattern generation and classification, in a two-stage process. In the first phase, a set of sequential patterns are generated from multi-dimensional sequence data. We proposes a novel method for inducing multi-dimensional sequential patterns with the use of Hellinger measure. The importance of each sequential pattern is also calculated. In the second phase, the generated sequential patterns are used for classifying multi-dimensional sequence data. A number of theorems are proposed to reduce the computational complexity of generating sequential patterns. The proposed method is tested on some synthesized sequence databases.