Research On Weibo Topic Tracking Based On Evolution Of Feature-items Ontology

Microblog as a new kind of online social media platform, developing rapidly in recent years and used by more and more people, has become a important platform of the general public access to information and opinions and increasingly become the most important source of information in the public opinion, rapid and timely access to weibo hot information became the urgent needs of the people. So it is necessary to introduce the topic tracking technology which allows users to quickly accurately grasp the microblog hot topic. Traditional topic tracking technology mainly uses the keyword model which has big data dimension and is not accurate. So this paper track topic through extracting the ontology of the most relevant information. at the same time for the flaws of the independent keywords on the topic semantic representation, this paper puts forward the concept of feature items ontology and applied to the topic tracking.This paper first introduces the topic tracking related basic theory, details the topic tracking process, the key technologies involved and the shortage of the existing topic tracking methods. at the same time this paper introduces ontology knowledge, and analyzes the characteristics of the ontology and the concept of the ontology evolution. Then aiming at the shortcomings of the existing topic tracking methods, analyses the semantic features of weibo report, put forward the topic representation based on ontology, and automatically build the initial topic ontology. the characteristics of the single item of the initial ontology can not fully express the topic semantic information, and thus puts forward the concept of feature items ontology. This paper tracks microblog topic by automatically extracting topic characteristics of terms to describe the events in this topic information and the ontology evolution. Finally, this paper design the corresponding experiment system, the experiment is carried out to verify the proposed approach.The specific innovations are as follows:(1)According to the semantic characteristics of microblog this paper puts forward the feature items ontology to represent the microblogging topic. the Feature-Items ontology can be constructted and evoluated automatically.(2)putting forward topic tracking algorithm based on the ontology evolution of microblogging and according to the features of the microblogging topic designing the corresponding weights and similarity calculation method.