Analysis of the Key Factors of User Behavior Measurement in Unexpected Events based on Feature Extraction and Selection

Abstract

Artificial intelligence is important for people’s daily life. In recent years, short video platforms have risen rapidly, and TikTok APP is one of the most outstanding among them. AI tools are applied to support their big data recommendation algorithm. In the era in which everyone is using the Internet, the public can often find the reflection of all kinds of attitudes and behaviors when unexpected events occur, and this influence will inevitably become more and more important as well as draw more and more attention. Therefore, taking the key factors of the user's behavior on the TikTok platform as the research object, this thesis takes “3.21 China Eastern Airlines Passenger Plane Accident” event as the background to extend the topic. Then, the existing data of TikTok platform such as Cicada Mama is used to capture and analyze the key factors of users’ behavior in the backdrop of unexpected events, such as user attention, user influence, the heat of the event itself. Next, according to AI methods of statistics-related data, the calculation method for each factor is proposed. Finally, the analytic hierarchy process is used to establish the method of key factor measurement, to analyze the weight of each factor in the emergency.