User Behavior Analysis Of Mobile Big Data Based On Hadoop

The world has entered the era of big data.For telecom operators,with the development of mobile Internet technology,a large amount of user data has been acquired.Through the analysis of these data,including the user's location distribution,Internet trends,consumer behavior analysis,help communications operators to improve operational models and meet user needs.This paper focuses on the analysis of large data processing method of mobile research,using Hadoop mobile big data platform for storage and processing of mobile data,and the background of the case based on the actual project,we propose an improved K-means algorithm,user behavior analysis of users,user traffic flow,verify the effectiveness of the analysis method of mobile user behavior.The main work of this paper includes:First,using Hive and Arc GIS technology based on Hubei mobile Hadoop platform and user data resources,according to the "2016 Wuhan marathon big data report" project,the preprocessing and visualization analysis of user data;the mobile user mobility and user traffic in the correlation analysis,the results show that the user mobility and the user traffic is directly proportional.Second,aiming at the defect of K-means is easy to get the local optimal solution,put forward the improvement method to choose the maximum density of some initial points by the Euclidean distance,the improved algorithm is applied to the mobile user behavior data flow clustering analysis,The experimental results show that the improved K-means algorithm can improve the accuracy of the algorithm withoutaffecting the efficiency.