Discussion on user behavior analysis methods of IP Networks

 

What is privacy? If I want to provide good services for you, can you let me know your behavior or habits?

...........................................

 

Discussion on user behavior analysis methods of IP Networks

(08:48:27)

China Telecom is in a period of transition to modern integrated information service providers. Implementing precise management is one of the key measures to achieve China Telecom's strategic transformation. The analysis of user behavior on IP networks is one of the tasks of Precise Management of Internet services and the prerequisite and guarantee for targeted marketing for users. Through in-depth analysis of user behavior on IP networks, you can understand the real needs of users, make full use of network resources, develop attractive value-added services, and increase user dependence and loyalty on networks, and then build a better profit model. Analysis Requirements of user behavior on IP Networks The user behavior analysis requirements of IP networks are diverse, and different business departments have different priorities. The business department focuses on user value analysis and value-added analysis to provide a basis for formulating corresponding marketing strategies and providing new telecom value-added services, such as users' surfing time habits and hot network businesses; the Operation and Maintenance Department focuses on the normal operation of the network, such as user satisfaction analysis and abnormal traffic. The Planning and Construction Department focuses on network traffic models and business development trends related to network planning. In addition, user behavior analysis can also provide valuable analysis reports, such as hot business development analysis and major account network analysis reports. In order to fully explore network operation data and systematically analyze user behavior of an IP network, user behavior analysis needs can be sorted in three dimensions: User, business, and traffic. Analysis Model of IP network user behavior User Behavior Analysis on IP networks is a dynamic and complex process. First, as users' Internet habits and consumption habits evolve with the development of the Internet, the demand for user behavior analysis is constantly changing. Second, users' online habits may change due to the network control and Business Development Strategies of telecom operators and information providers. Thirdly, user behavior analysis sometimes does not have a clear vision, the actual analysis requirements may be gradually improved in the data analysis and processing process. Therefore, a complete set of models is required for user behavior analysis to adapt to the dynamic and iterative correction analysis process. The entire analysis model can be divided into four parts: requirement proposal, business data collection, data analysis, and result evaluation. Demand proposal: The analysis requirements can be raised from the user, business, and traffic dimensions. For each analysis item, a user range needs to be determined. Demands are raised based on business needs, and new analysis projects can be found during data collection and analysis. Data collection: On the one hand, data collection technologies such as probe detection, Netflow Collection, and DNS access statistics must be used to obtain data from user access networks. On the other hand, user attribute data and value data, business data of the enterprise and competitors, and other business data need to be obtained from the business system or through intelligence analysis. Data Analysis: Based on the analysis requirements, relevant data filtering, preprocessing, and comprehensive data analysis and processing procedures are developed based on network data, user data, and business data to obtain valuable analysis results, and displayed in an accurate and intuitive manner. Result Evaluation: Compare the analysis requirements and analysis results to evaluate the analysis results. Several parts of the analysis model are linked and have the characteristics of information feedback and gradual improvement. Data collection policies are formulated based on analysis needs. At the same time, new or more specific analysis needs are guided through data collection and preliminary analysis; the analysis results provide assistance to the business department on the basis of meeting the analysis requirements. The business department should improve the analysis content based on the analysis results and develop new analysis requirements. Several methods of user behavior analysis on IP Networks In addition to analysis requirements, data analysis methods are also the key to user behavior analysis. The user behavior analysis process of an IP network is actually a data mining process that obtains valuable information from massive data. Therefore, user behavior analysis can refer to some methods in Data Mining and analysis disciplines. Based on the characteristics of user behavior analysis, user behavior analysis can include user Feature Analysis, association analysis, classification and prediction, exception analysis, Top N analysis, and other analysis methods. User Feature Analysis User feature analysis refers to identifying the behavioral characteristics of various users. User feature analysis is a prerequisite for targeted marketing. For example, by comparing the usage characteristics of users of different tariff levels, you can obtain the characteristics of high-paying users accessing the internet, find the target users, and transfer users to high-paying users through business guidance and tariff policies. For example, by analyzing the characteristics of the business types used by government and enterprise customers, we can marketing business pilot products to major customers. Association Analysis Association analysis discovers association rules. In user behavior analysis, You can associate a user's habits with other habits, or perform Association Analysis on users' consumption habits and network habits, you can also analyze the association between users' natural attributes, such as age, gender, and occupation, and their network habits. Classification and Prediction The classification technology can be used to classify users into a specific class. For example, you can analyze a large number of users to obtain the online behavior characteristics of the user before the user is detached, to form a disassemble user model, and then compare it with the model, issue a disassemble warning to users with similar behavior characteristics to provide a basis for the marketing department to take customer retention measures. Exception Analysis An exception Analysis for IP network users is an analysis of abnormal network traffic such as hacker attacks, network viruses, and spam. Another exception analysis is to analyze individuals with abnormal behavior characteristics. For example, when a leased line user shows that the traffic has the characteristics of a public customer, and the traffic is far beyond the potential traffic of the leased line user, the leased line user may be operating a public customer. Topn Analysis Topn analysis is a common method for user behavior analysis. During analysis on a certain type of users, the topn users have the greatest impact on the features of such users, and the corresponding policies and measures taken for the topn users are also the most effective methods. Therefore, in user behavior analysis, the top user needs to be further analyzed. In short, IP network user behavior analysis is a comprehensive analysis for users, and its research and application are constantly developing and improving. I believe it will play an increasingly important role in Internet operation.