The Research And Application Of User Feature Discovery Based On The Social Networks

With the cyber-revolution caused by the development of information technology,the Internet is transfering from Web 1.0 Era,in which users search for information from the internet,to Web2.0 Era,in which users search for and also create information in the internet.As the typical application in Web2.0 Era,Social Network Services are quickly getting popular.In fact,the development of SNS is transfering the information management from offline to online,making the large amount of content in virtual networks.These information gives each node personality,and owns great potential business value.Based on this background,this paper studies the problem of user feature discovery based on social networks.Firstly,we build the user feature discovery model,which includes two types of problem:user feature discovery with tags and without tags.The algorithms used in this model include classification,clustering,graph mining and text mining.Then,we conduct the model in different real situations,including depression detection as the discovery with tags,ageing social community and user interest as the discovery without tags.The experiment result based on Sina Micro-blogs shows that the algorithms can complete the problems,and the accuracies are all above 70%.It proves that the solution in this paper is reasonable,and it can ensure the accuracy of discovery result,thus can meet the need in user feature discovery in social networks.At last,this paper develops a web application based on the model and algorithms,which aims at sentiment analysis and depression detection.we introduce its architecture,development environment,conduct introduction and performance analysis.