Approach to Recover CSGM Method with Higher Accuracy and Less Memory Consumption using Web Log Mining
Shrivastva Neeraj, Lodhi Singh Swati · 2012
Sequential pattern mining is an important mining technique which discovers closed frequent sub sequence from a sequence database. However it is very difficult as it generates explosive number of sub sequence in candidate generator and test approach. Previous sequential pattern mining algorithm closed sequence-sequence generator mining (CSGM) mine full set of frequent sub sequence satisfying a min_sup and max_sup threshold in sequence database. This algorithm is not suitable for datasets that are too dense or too sparse, which is prohibitively expensive in both time and space. In this paper we analyze the existing methods of sequential pattern mining and after analysis we propose an enhance algorithm for sequential pattern mining. Thus the main purpose this method is aiming to solve is to develop new techniques based on the closure concept for effectively and efficiently discovering non-redundant sequential association rules from sequential datasets with higher accuracy, less memory and time consumption. After performance analysis we use modified algorithm for mining useful data from web log. Keyword : Sequential pattern mining, CSGM, web mining.