Mining Sequential Pattern With Synchronous And Asynchronous Periodic Time Stamp Using Hash Based Algorithm

Sequential Pattern Mining (SPM) is considered as an interesting data mining problem conducted in research works. In SPM, the time-series pattern can be used as a tool for identifying the behavior of the patterns while mining sequential pattern types of data. Mining time series patterns in temporal dataset acts as an imperative function in data mining and knowledge detection mechanisms. Several authors have been focused on their research on mining time series pattern with synchronous and asynchronous periodic databases. But the downside of the researches is that due to the involvement of random noise and disruption, some periodic patterns were not recognized. So, to enhance the sequential pattern mining with both the synchronous and asynchronous periodic timestamp, in this work, we deploy a new technique termed hash based algorithm. The hash based algorithm finds all maximal complex patterns in a single step devoid of mining distinct event and various events patterns employing only one sequential dataset search. It introduces sequential pattern for varied periods and generate frequent pattern and candidate maximum pattern for an effective mining. An experimental evaluation is carried out with the datasets retrieved from the UCI repository used to evaluate the performance in terms of accuracy and scalability using hash-based algorithm to that of used to evaluate the performance in terms of accuracy and scalability using hash-based algorithm to that of GSP for asynchronous periodic pattern mining for asynchronous periodic pattern mining.