Mining serial positioning episode rules from event sequences and batch free top k sequential pattern discovery for data streams

Abstract

The Frequent Episode Mining (FEM) is a challenging problem to identify frequent episodes from a sequence database where the rules discovered from episodes are used for prediction and decision making. An episode rule mining to extract useful and important patterns or episodes from large event sequences represent the temporal implication of associating the antecedent and consequent episodes. The work presented here has four major contributions to mine frequent episodes from event sequences and sequential pattern discovery in streaming data. The first research contribution deals with mining of serial episode rules using natural exponent inertia weight based particle swarm optimization algorithm from event sequences. The second work corresponds to the mining of serial positioning episode rules from event sequences using natural exponent inertia weight based swallow swarm optimization algorithm. In the third contribution, frequent serial episode rules are mined using forward and backward search technique from event sequential data. An adaptive hierarchical clustering and batch-free top-k sequential pattern mining from data streams is the fourth major contribution of this research work. newlineThe episode rules need to be mined with precise and serial based rule mining considering the temporal factor, so that, the occurrence time of the consequent is specified in contrast to the traditional episode rule mining. In the first proposed work, the fixed-gap episodes are generated with specific time constraints, and mining such episodes from whole sequence where the time span between any two events is a constant, are utilized to improve the system s performance. The existing technique for mining precise-positioning episode rules from event sequences, mines serial episodes resulting in enormous memory consumption newline