IPTV Real-time Analysis System Based On Big Data

In recent decades,with the continuous promotion of business integration in broadcasting,television and telecommunications,the development of new media services of television stations in various places has been promoted.As an important consumer product in the era of smart homes with the Internet of Things,smart TV and digital TV technologies have also been rapidly developed,and at the same time,the TV programs people watch have become increasingly abundant.The analysis system for IPTV is also extremely important.However,in the existing analysis system,there are many problems,such as the transmission of massive data and the simple use of users.Several types of data have very limited use of user viewing behavior data.At the same time,the system did not take into account the low real-time performance caused by the expansion of the data volume when the system was constructed,and it has not been well resolved.In order to solve the problems of low data collection efficiency and inaccurate user behavior analysis in the existing IPTV analysis system,an IPTV real-time analysis system is proposed.In the existing analysis system,how to efficiently and real-time data collection needs to be solved urgently.In the IPTV scenario,user data is extremely rich and huge.Flume is widely used in data collection.However,this article compares existing solutions to the deficiencies of the Flume data collection framework.An improved Flume framework is proposed to realize a real-time data collection framework.This method combines the zero-copy technology in the Linux environment with Flume's Source to achieve more efficient data collection.In order to make full use of the advantages of Flume's two channels,it is proposed to adopt a composite channel.The system can be used independently according to the current situation.Choosing the appropriate Channel can improve the efficiency of data collection.In the analysis system,data analysis is the most important link.This article analyzes the three services of IPTV users and the combination of TV programs and other data,and proposes an improved GSA algorithm SVM parameter optimization user behavior analysis model,the model uses the optimized SVM algorithm to use the IPTV service and the program watched under the service.On the basis of time distribution,the user's viewing behavior characteristics during workdays and breaks are analyzed,and the analysis results can be fed back to the enterprise.Finally,this paper implements an IPTV real-time analysis system on the basis of big data clusters,and completes the development of various analysis services in the system.At the same time,the system is functionally tested.The final result shows that the system can better analyze user behaviors.At the same time,it can provide favorable technical support for the future strategy formulation of the telecommunications platform.