Web Usage Mining in Online Social Network

The web content in present scenario is mainly comprised of Social media systems such as blogs, photo and link sharing sites and on-line forums. . Web Usage Mining is the application of data mining techniques in the field of social networks to discover exciting usage patterns from SNS data and to serve the needs of SNS applications in a better manner. The major use of web usage mining techniques has been confined to web logs. But we use the same techniques to discover the hidden relationships between different nodes within a network or across networks. The most commonly used Apriori algorithm had a major disadvantage of performing multiple database scans for candidate set generation. The FP Tree structure solved this problem by restricting the database scans to two times. But the FP growth algorithm was very complicated and time consuming since it recursively created trees at every step during frequent item-set generation. The FP-Split algorithm further improved candidate set generation by doing a single scan of the date for candidate generation. The Apriori growth algorithm when used for mining the FP Tree performed better in frequent item-set generation. So we now propose a hybrid technique for web usage mining using FP Split Tree and Apriori Growth algorithm in the field of social network mining to perform analysis of user behavior on these sites. This can be advantageous while creating communities within a network, defining trust relationships, user behavior analysis and for friend suggestions. Keywords—Apriori Growth, FP Split, SNS, frequent patterns.