Hybrid Technique for Frequent Pattern Extraction from Sequential Database

Data mining has became a familiar tool for mining stored value from the large scale databases that are known as Sequential Database. These databases has large number of itemsets that can arrive frequently and sequentially, it can also predict the users behaviors. The evaluation of user behavior is done by using Sequential pattern mining where the frequent patterns extracted with several limitations. Even the previous sequential pattern techniques used some limitations to extract those frequent patterns but these techniques does not generated the more reliable patterns .Thus, it is very complex to the decision makers for evaluation of user behavior. In this paper, to solve this problem a technique called hybrid pattern is used which has both time based limitation and space limitation and it is used to extract more feasible pattern from sequential database. Initially, the space limitation is applied to break the sequential database using the maximum and minimum threshold values. To this end, the time based limitation is applied to extract more feasible patterns where a bury-time arrival rate is computed to extract the reliable patterns.