Empirical Study Of Zipf's Law In Microblog Hot Topics

With the continuous development of internet technology, various mobile platforms are widely used and instant messaging clients are increasingly becoming more and more. Microblog, a social network platform, is vigorously sought after by the internet users ever since it came to the world. It greatly satisfies the non-contact social needs for the internet users, not only because its characteristics of real-time, originality and flexibility, but also because it allows the users to choose personalized following up and interaction; thus, microblog rapidly becomes an important application in social network platforms. The current users of microblog have reached the level of one hundred million; such a large group makes lots of microblog hot topics wide-ranging spread and perfect through grading transmission. Microblog has become an important bridge and link for information communication. Many events are first reported by microblog; in this micro-era of mass communication, the research on microblog hot topics has become a crucial issue. Therefore, automatically processing of massive microblog information by computer and study on microblog topics spread pattern have significance on understanding social network structure, providing insight into public opinion and grasping public opinion trends.In this study, in order to obtain the relevant data from massive microblog topic information, the method of microblog platform open API combined with web search is used, especially the tencent microblog platform open API, including Oauth2.0 authorization, database storage and effective information merging and classification, which makes access to data and filters effective information efficiently. In the research, Sina microblog and Tencent microblog, which have most domestic users, are selected as objects to investigate the characteristic analysis of topic data from single and multi-topic sources. In the single-topic research, 11 representative topics are selected and analyzed. In the multiple-topic research, totally 1731 hot topics are selected and analyzed,which are chosen based on the Sina microblog hot list and Tencent daily hot topic recommended. Through data screening and function fitting, the final conclusion is reached that the frequency and rank of microblog hot topics fit Zipf's law. In summary, this study mainly works on the following three jobs: analysis of microblog user network structure and topic transmission cycle; study on how to obtain microblog information based on microblog platform open API; and finally, verification of topic data in line with Zipf's law through analysis of sudden single- and multi- hot topic data.