Frequent Itemset Mining over data streams with unbounded and continuous flow of data is playing an active role in data mining towards knowledge discovery. Sliding Window (SW) is recognized as standard technique for deriving knowledge from such data streams. Due to unbounded data, the window size is fixed, and window moves from one to another window by ignoring few transactional data and replaced with new transactional data to derive up-to-date knowledge.It is not wise idea to ignore itemsets that are in the past window since itemsets effect remains for some time. Computing and keeping all frequent itemsets is not a wise idea due to huge number of frequent itemsets and redundancy in a result.Closed Itemset (CI) Mining is an alternative solution to reduce the redundancy among FI’s. Hence Closed Itemset Mining(CIM) using sliding windowis interesting than complete frequent itemsets.To retain the influence of the itemsets ignored in past portion, Weight Decrease concept is used. Here, we propose SW-CFIW algorithm for mining CI’s. This algorithms uses the prefix tree to keep itemsets and transaction information in a condensed form. Each node of the tree represents the weighted frequent itemsets. Experiments are conducted onstandard datasets are showing that our approaches are giving good performance.
Volume 11 | Issue 7
Pages: 992-1003