CONSTRUCTION OF INDIVIDUAL COGNITIVE MODEL IN THE CONTEXT OF EMOTION AND AFFECTIVE DISORDER
Weidong Huang,Qian Wang,Xinkai Yao,Junyan Tian
DOI: https://doi.org/10.1093/ijnp/pyac032.056
2022-07-01
Abstract:Abstract Background With the development of various information technologies and the popularization of the Internet, great changes have taken place in human life. In particular, social media has become an important channel for people to obtain information and generate event awareness, attitude and emotional tendency. Personal real-time comments on social media have a potential role in guiding public opinion and guiding network public opinion. As one of the most representative online social media, microblog contains a lot of public emotional information. The heterogeneity of microblog users' personal characteristics such as age, work and education leads to their different views, attitudes and values, making their views more representative and universal. Emotional analysis of these information is of great significance for us to understand the spread of public opinion events. By constructing a personal cognitive emotion classification model based on the optimized distributed gradient enhancement library, the emotion classification of microblog short text is realized. The model mainly includes text data collection and preprocessing, text feature extraction based on word embedding model word2vec, high fitting emotion classification model established through training corpus, and emotional tendency judgment according to the emotional tendency results obtained from the model. This paper takes 5 million microblog data and the data set released by CCF international natural language processing and Chinese Computing Conference as the word vector training corpus of word2vec to train the emotion classification model. Research Objects and Methods In order to investigate the semantic analysis of emotion, we formed an open-ended questionnaire by consulting relevant literature, sorted out the recovered effective scales, and conducted data semantic analysis. On this basis, combined with the previous theoretical concepts, the preliminary measurement table of emotional information is compiled; Secondly, the preliminary test is carried out according to the items of the scale, and the data are analyzed through item analysis, exploratory factor analysis and reliability test of the scale; Finally, the formal scale is compiled, tested and recovered, and verified by confirmatory factor analysis and reliability and validity test. In order to test the relationship between emotional information and emotional regulation, the cognitive bias questionnaire (CBQ) was used to measure the negative cognitive bias assumed to be related to depression. The scale measures two dimensions: (1) depression and (2) cognitive distortion. “Depression” expressed by CBQ refers to depressive emotion, or bad mood, rather than a depressive syndrome containing all relevant symptoms. Cognitive distortion is defined as reasoning that is clearly incorrect in terms of known facts. CBQ attempts to evaluate the specific cognitive distortions proposed by Beck (19671970), such as over generalization, out of context, arbitrary and speculative reasoning and judgment, ignoring advantages or good results and exaggerating disadvantages or bad outcomes. The scale describes six situations common to college students or psychiatric patients, three for interpersonal relationships and three for self-achievement. After each situation, three or four questions are put forward. These questions represent four possible combinations of the two dimensions of depression and distortion: depression non distortion, depression distortion, non depression non distortion and non depression distortion. Subjects were asked to answer how they experienced when they were in this situation. The score is to compare the scores of the four combinations of depression and distortion respectively, and the score range is 0-23 points. At the same time, independent sample t-test and analysis of variance were carried out with college students' gender, grade, family location, major, one-child situation, student cadres, family status and parental education as independent variables. Whether it is a student cadre or not has a certain impact on the emotional regulation and self-efficacy of college students. Results The improved model overcomes the shortcomings that the general methods can not explain the context and capture the semantic relationship of the context, and improves the efficiency and accuracy of emotion based classification. The effectiveness of different models can be evaluated according to confusion matrix or ROC curve. In this paper, the classification effect of xgboost emotion classification model is obviously better than that of SVM classification model. Conclusion Our empirical results show that most netizens hold a negative attitude towards yumengyao wrestling competition. Due to its limitations, this paper only uses xgboost model to study the text emotion classification. In the future, deep neural network can more accurately understand the characteristics of microblog short text, realize higher precision emotion analysis, and help practitioners and decision makers comprehensively and effectively identify the public's views and attitudes towards national policies and current events. This paper is of great significance to the real-time public opinion analysis in the network social media environment, and emphasizes the necessity of promoting the establishment of network public opinion early warning mechanism. Acknowledgements Funded by the National Natural Science Foundation of China under Grant No. 71671093. The authors would like to thank other researchers at Nanjing University of Posts and Telecommunications. Supported by the Jiangsu Provincial Department of education, the key research base of philosophy and social sciences.
pharmacology & pharmacy,neurosciences,psychiatry,clinical neurology