Optimal pattern search for database systems

Many time series database applications require processing and analyzing database sequences where the focus is on finding the patterns and trends in sequence data. In this paper, the approach carried out is based on the observation that finding sequential patterns in databases is somehow similar to searching for a phrase in the text. However instead of searching for sequence of letters usually from finite alphabet, we search for sequence of tuple with rich structure and infinite possibilities from both ends of the sequential database. We take the algorithm of Boyer moore used for text and generalize it to search for complex sequence patterns of interest in a given database by exploiting the logical interdependencies between the elements of a sequential pattern having constraints.