The Application Of Hadoop Based Data Mining In Telecom Customer Analysis

With the development of information society,the amount of data that people can gather is becoming larger and larger.How to find hiden useful information from the data we gathered is now becoming a main concern of people.In order to solve this problem,people designed the Hadoop platform and developed data mining technique.Hadoop is a distributed platform designed to process massive data.Data mining is a technique which can dig out hiden information from data.If we combine these two techniques together,massive data mining will not be a problem.Based on the study of Hadoop and data mining,this paper make a deep research on the possibility of the combination of Hadoop and data mining.The main contribution of this paper includes:1>Implementation of the Kmeans algorithm using Hadoop MapReduce framework.Accomplished the cluster analysis of telecom customer using this implementation.Process a simple analysis of the cluster result referring the product provided by the telecom operators.2>Implementation of the KNN algorithm using Hadoop MapReduce framework.Accomplished a classifier using this implementation.Process a simple analysis of this classifier.3>Bring out a recommendation system based on the collaborative filtering algorithm.This recommendation system was based on the characteristic of telecom customers.It calculates similar customers based on the customer’s behavior instead of on the product the customer bought.