Algorithm for Mining Sequential Pattern in Time Series Data

Mining sequential pattern in time series data is broadly used in a variety of areas in order to make a prediction, and an appropriate model should be established before the prediction can be done, therefore, the way how to mine out time series pattern from time series database becomes extremely important. Based on data of the time series database, this paper presents a new frequent time series pattern mining algorithm, which constructs a tree-projection at first, then uses priority depth strategy to traversal the tree-projection in order to mine out all the longest frequent patterns, the paper has descripted the new algorithm with pseudo code in detail. Experimental results demonstrate that this algorithm has mined out the frequent series, which meets the real-time restraints successfully. Furthermore, under the same condition and different support situation, the new algorithm has obtained the same ruleset as the traditional AprioriAll method but more effective performance.