A Review: Frequent Pattern Mining Techniques in Static and Stream Data Environment

Objectives: This paper focuses on various frequent pattern mining techniques, their challenges involved in static as well as stream data environment. Analysis: Information, the most precious asset is most of the times hidden inside heaps of raw data. It is required to be polished to retrieve information, converted into a form which can be analyzed for making decisions. Many research papers were read for the understanding of various techniques that mine frequent item sets either in static or stream data environments. Findings: This paper summarizes all the popular techniques available with us for frequent item set mining and provides some suggestions for optimizing them. The issues for static mining just take into account the time and space complexities but stream data mining is much more complex and challenging as compared to static. The problems like concept drifting, nature of data, its processing model, inadequacy, size, its retention in warehouses etc. are also required to be taken into account while working with real time data. Improvements: The algorithms for stream data mining should be incremental and resource adaptive in nature so that they can handle the change and adjust their processing parameters according to the availability of resources respectively. The available resources which will be very helpful in shared environments (where resources are shared by multiple processes).