Objectively Measure Subjective Feelings

In social management,economic decision-making,and other problems,we usually need to figure out the subjective feelings of an object or a group,such as interests,emotions,and preferences.Traditionally,the most common method is the questionnaire.However,besides design difficulty and heavy workload,this method has been increasingly questioned in its practicality.In the era of big data,with the development of information technology and the emergence of social media,there are now new ideas and possibilities for the acquisition and measurement of subjective feelings.In the context of objective measurement of subjective feelings,this thesis primarily focuses on inferring social anxiety psychological status based on social media.This thesis proposes a new framework for social anxiety inference.Different from traditional identification methods,this framework directly collects objective information from social media,determines relevant features,and then,builds an inference model.Combining psychology,linguistics,and other domain knowledge,this thesis proposes a series of features such as emotional trend,emotional index,and interactive index,to describe and distinguish users.Compared with the current psychological status research based on social media,this thesis not only focuses on the empirical results but also pays great attention to the interpretability behind the model.Moreover,prior knowledge plays an important role.This work analyzes the fine-grained emotional features of users and constructs a multivariate statistical analysis model for social anxiety inference,based on the mapping relationship between the text and the feature,the relationship between the feature and the judgment target,and the cross-relationship between the features.Finally,this thesis uses real-world data to evaluate.The results demonstrate that the proposed model is distinguishable and can well infer users’ social anxiety status through social media information,and also illustrates the feasibility of objective measurement of subjective feelings.