A journey of frequent itemset mining techniques

Information technology has become integral part of every other area of work. People have become smart and connected to the world all the time. Industries are using databases to store their data. So a huge amount of data is produced every second. This data can be very useful to companies to make business related decisions and to enhance their business by making customer luring policies. So, there is need of techniques, which can extract useful information from such huge amount of data and present that information to the user in an efficient and understandable format. A need of such information was first felt in 1993 when Agarwal studied the supermarket transactions dataset i.e. to examine customer behavior in terms of purchased products. After that many researchers have worked in the area of frequent item set mining and hundreds of algorithms have been proposed. There are basically two types of algorithms, Apriori based and FP-tree based. So there are a lot of techniques in the literature, some of which are reviewed in this paper in detail. Techniques are, Apriori, Partioning approach, Bit-Table approach, Interval Intersection algorithm, Vertical map-reduce and Parallel fast update. A comparison study of these approaches is also presented with their respective advantages and disadvantages.