Analysis of Enterprise User Behavior on Hadoop

Abstract

In the past few years, the network is an essential part in human's life. More and more people choose shopping, chatting and video online. All the online activities produce tons of data which contain our behavior characteristics. Traditional data analyses just focus on one kind of feature, while they neglect the behavior information generated by merging multiple features. Based on the basic data set, this paper provides a new method for analyzing data, called Compound Session. After acquiring the Compound Session data, we continue to process data from the perspective of enterprise user and produce three types of analysis tables. Due to huge amounts of data from more than two thousand enterprises, we propose to process data on the cloud computing platform called Hadoop.