Application Of Improved Apriori Algorithm In WeChat Hotspots Analysis

With the development of Internet technology, new media based on mobile terminals are emerging. WeChat now is highly recognized by domestic and foreign users, due to the features of miniaturized content, fast spread speed, supporting real-time intercom, and integrating social, communication and platform. In WeChat hot phenomenon often occurs, namely some topic become the hot talk and visit in a short period of time. Positive hot phenomenon has good propagation effects, negative ones will affect the harmonious network environment. Therefore, the analysis of public opinion based on the hot topic of WeChat has become a new research subject.In this paper, we associate the analysis of WeChat hot topics with the rule mining technology, use the association rule algorithm to extract the valuable information hiding behind WeChat data. That is to say, according to the hot words involved in WeChat hot topics, explore the strong association rules between hot words, analyze the phrase with strong representative in hot topics. In order to improve the processing efficiency of WeChat data and guarantee the analysis quality of data, we have made some improvements on the Apriori algorithm of association rules. The main idea of improved algorithm is to integrate upper triangular support matrix with transaction binary vector, improve the speed of generating frequent 2-itmesets, which aimed at improving the execution efficiency of algorithm and protecting the effectiveness of strong association rules among hot words at the same time.The experimental results show that, based on the same set of experimental data, the improved algorithm has better performance than the traditional Apriori algorithm, under the different conditions of transactions counts, support threshold and item numbers. Meanwhile, applying the improved algorithm to WeChat data can effectively extract strong association rules among hot words in hot issues, which has realized the combination of WeChat hotspots analysis and association rules mining technology, achieved the desired results.Based on the above theoretical analysis and experiment shows that this project has laid the theoretical and technical foundation for the future in-depth study of WeChat public opinion analysis and hotspots detection.