Time Series Analysis Of Hot Words In Social Network

With the rapid development of technology, the Internet has become a great media of information dissemination. Social network has obviously changed the life style of common web users, and it is more and more relied on to gain and share information because of the convenience. Therefore it becomes a modus of online mouth branding. Hot words in social network,not only affect people’s views of real life, but also reflect social opinion in different ways. So the research of hot words has very important practical significance, and the trend of hot words has become a hot study subject.Because of the diversity and randomness of hot content in social network, this paper focuses on time series analysis to study propagation rules of hot words. Fuzzy mathematics theory is introduced, and the algorithm of fuzzy set partition is proposed to classify the data of time series.The algorithm of fuzzy trend analysis is also proposed, reflects the process of transmission in social network. In this paper, the temporal patterns of hot words are introduced based on fuzzy time series theory, which uses the soft computing method to improve accuracy of classification. Fuzzy membership functions are applied to determine fuzzy sets, which increase flexibility and diversity of time series classification schemes. Based on fuzzy inference rules, specific trends of time series values are schemed to enhance accuracy of time series prediction.In this paper, based on fuzzy mathematics theory, an algorithm for time series analysis is proposed. And hot words of Sina Weibo are selected as experimental data. Compared with traditional method of time series prediction, experimental results show that the algorithm of fuzzy time series analysis performs well as for capturing the rise and fall trends, and achieves higher accuracy of time series prediction of hot words in social network.