| With the "Internet+",has become a popular concept,Businesses are closely integrated with the Internet up.It produces vast amounts of data.Currently we have been in the era of data,information overload.However,The user is difficult to find what she is interested in,facing massive information.While Internet companys analyze user interest,to provide users with better products,better services.Log data record the user’s behavior.How to analyze these massive log data generated by these products becomes very valuable.In the current research on data mining,the mining on association rules is a very important research area.The research involves many fields of information,health care,transportation,finance,etc.In the association rules Algorithms,the Apriori algorithm proposed by R.Agrawal and R.Srikant in 1994 is the most important and classic.The core idea is to generate a candidate set and plot closed down two-stage detection to mining frequent item sets.This paper focuses on the research on user massive data processing algorithm,which mainly include the following contents:(1)The usage of Apriori algorithm in user video recommendation system.In this research,we acquisite and process a famous internet company’s massive log data,through mining the degree of association between different videos,and to recommend more relevant video to users.(2)The research of improvement and application on Apriori algorithm,due to open database frequently and a large number of frequent itemsets generated leaving bad time performance.In this paper,we propose some optimization strategies of frequent item set encoding on the basis of a large number of log data analysis.(3)The research of Apriori Based on MapReduce.Faced with TB,even PB-class massive,log data,stand-alone computing is difficult to meet the needs of large data mining.In this study,we improved Apriori algorithm based on Map-Reduce,so that it can be applied to data mining based on Cloud Computing... |
PDF Full Text Request