Analysis of User Attention Behavior and Its Driving Factors on WeChat Public Platform

In recent years, with the development of new media, such as WeChat Public Platform and Tiktok have become an important source of news and knowledge for the public. All kinds of official account have sprung up one after another. How to operate them to attract more public attention, so as to better serve the science popularization and knowledge dissemination of public welfare units, is the problem faced at this stage. This paper takes the WeChat Public Platform of Hubei Institute of Geoscience as an example, based on the data of the platform, uses XGBoost algorithm to evaluate the factors that attract the public and analyze the driving factors of their attention behavior from several aspects, such as the number of posts, article types, layout design and interaction. The following conclusions can be drawn: ① The public attention behavior factor evaluation model based on XGBoost algorithm integrates machine learning technology and data mining ideas and achieves good evaluation results; ② According to the ranking of feature importance, the type of article has the greatest impact on public attention behavior, followed by interactivity the evaluation results show that the contents of the WeChat public platform of the Hubei Institute of Geoscience involve popular geological knowledge, the latest research results, project activities and other contents, taking into account the booking consultation and link functions of the Hubei Geological Museum, which can effectively serve the public. The research results provide decision-making basis for the next operation of the public platform.