Automatic Analysis Of Explicit And Implicit Social Attitudes In Social Media Language

In social psychology research,the Self-Report Assessment and the Implicit Association Test are used to measure the explicit and implicit attitudes of groups,respectively.However,both methods require the active cooperation of participants,and can only be performed on a small scale groups' attitude measurement.The comments made by users in social media can reflect their attitudes,but current text attitude analysis methods do not distinguish between explicit and implicit attitudes of groups.Inspired by social psychology attitude measurement methods,this paper aims to study an automated method for measuring and distinguishing explicit attitudes and implicit attitudes of large-scale groups,and is no longer limited to psychological experimental measurements of small-scale groups.By linking social psychology measurement methods and language statements at a high level,and simulating experimental standards in Self-Report Assessment and Implicit Association Test through statements with specific language patterns in people's statements,thereby distinguishing between two groups that reflect explicit attitudes and implicit attitude language statements,and then use the sentiment analyzer to calculate the average sentiment score of the group's explicit attitude and implicit attitude statement set to understand and quantify the group's explicit attitude and implicit attitude.In the experiment,four different social media corpora(Chinese Wikipedia,English Wikipedia,Chinese Weibo,and English Twitter)were selected for related group attitude measurement experiments.A large number of experiments have verified the effectiveness of the attitude measurement method based on group language sentiment proposed in this paper.At the same time,this study also explores and analyzes the evolution,stability,and correlation of the group's explicit attitude bias and implicit attitude bias on the Chinese microblog corpus and the English Twitter corpus.