Association Rules Mining On Telecommunication Of The Recommendation About Similar Users

In the information era,data shows a trend of explosive growth.Walks of life today have paid attention to seek ways to extract valuable information from huge amounts of data.Especially the telecommunications industry produces hundreds of millions of data every day,and for such a huge database,operators,in order to improve marketing methods and quality of service,helping achieve the goal of enhancing market competitiveness and improving corporate profits,will focus on making good use of users' data and fully digging useful information.The significance of data mining is to provide operators with a new market situation analysis,and combine business model and data mining together to help operators make decisions easier and improve the market competitiveness.The paper combs the related literature and researches.It firstly introduces the clustering analysis and association mining of data mining detailedly.Clustering analysis and association analysis used in this article are based on the platform of Spark.After analyzing the basic China Unicom business of city A,k-means method is adopted first to classify users' basic business,and then digs the association rule of the users' ordered businesses.The users of China Unicom of city A will be classified according to its basic business consumption.After this,it may need to dig the associated mining rules for each class of users producted by clustering algorithm.The association analysis,which is used to analyze the correlation between different transactions is one important part of data mining.Besides support and confidence,the author also uses ascension and slightly improved,not generating candidates FP-Growth algorithm to find frequent items,and digs association rules according to the ordered businesses.By the end of the article the analysis results will be summarized.It concludes that users of China Unicom of city A at present are divided into three categories according to its basic business.It is divided into primary and middle school students and old groups,youth groups who love to surf the Internet and office workers.Then it will use association mining among these three kinds of users,and analyze the businesses which has been ordered.It is concluded that the primary and middle school students and elderly groups will be introduced basic businesses,youth groups will be introduced Internet businesses and office workers will be recommended group businesses.These advices will be offered to enterprises detailedly and reasonably.