Research On Affective Computing For Crowd Psychology

intelligent computing is a focal point of competition in the information industry and is being viewed as a key technology of the world in the 21 st century.Affective computing and crowd psychoanalysis are new directions that have received increasing attention in the fields of intelligent computing and mental health,combining cognitive science,psychology,mathematics and computer science.Therefore,crowd psychology-oriented affective computing is such a research hotspot of interdisciplinary research in mathematics,information science,and psychological science that is of great significance for the elucidation and interpretation of psychological changes in and public opinion analysis of crowds in society as well as for the optimization and promotion of the related AI applications.In the context psychological problems,this study investigated the techniques and methods related to affective computing based on the affective computing and big data technology.This paper focuses on the psycho-affective computing method and classifies it according to the data collection method.The psycho-affective computing method is mainly divided into two categories: survey method(indirect)and test method(direct).Therefore,this paper optimizes and improves two kinds of psychological and emotional computing methods,and takes primary and secondary school students and epidemic population as application examples to analyze and verify the effectiveness and practicability of the methods.Firstly,because elementary and middle school students cannot maintain concentration for such a long time that are unable to complete psychological instruments effectively,it is of great significance to study the simplification,i.e.,reducing measurement times while accounting for accuracy,of psychological scales for population cohorts using mental health measurements of elementary and middle school students as an example.Secondly,Secondly,due to the emergence of a large number of new words on social media,the traditional emotion dictionary expansion algorithm cannot quickly and effectively find new emotion words in the social media environment,nor can it judge the emotional tendency changes of early emotion words.Therefore,it is of great significance to study the expansion algorithm of emotion dictionary for analyzing the psychological and emotional changes of crowd.Finally,the calculation of the new found due to the expansion of emotional dictionary algorithm is very expensive,research and its matching high-performance parallel computing framework to improving the efficiency of the algorithm,reduce the running time,at the same time to build the corresponding large data analysis platform for the researchers to further explore the emotional calculation of psychological physiology meaning behind it to provide the reference.To sum up,this study explored the related issues of improving the calculation method of psychological emotions of primary and middle school students and COVID-19 population as examples.In short,using the mental health assessment of elementary and middle school students and an analysis of stress emotion trends for the population during the COVID-19 pandemic as a case study,this study aimed to simplify psychological scales and analyzed stress sentiment trend with following three categories.1.This study proposed a combined psychological scale simplification method that fuses multiple theories,based on which a simplified scale for the mental health assessment of elementary and middle school students and its corresponding selfadaptive computer measurement system was constructed.Specifically,we firstly proposed and implemented a combined psychological scale simplification method that integrates multiple theories based on a simplification algorithm to improve the mainstream scales.Secondly,we simplified and implemented a mental health assessment scale for elementary and middle school students that comprehensively assesses the mental state of elementary and middle school students from five aspects,i.e.,psychology,behavior,emotions,interpersonal,and environment.In combination with cohort data,we confirmed that the simplified scale obtained through the PSMU method simplified the existing scale items to the greatest extent while maintaining high reliability and validity.Thirdly,on the basis of the simplified scale for the evaluation of the mental health of elementary and middle school students,we developed and launched an adaptive measurement system to assess the mental health of these students.2.This study proposed an expansion framework for a sentiment dictionary,which helps us to analyze the crowd stress emotion trends.Specifically,we firstly developed a corpus data collection program through which more than 1.2 million corpus data were successfully collected.Secondly,by improving the construction method for mainstream sentiment dictionaries,we proposed and implemented an expansion framework for a sentiment dictionary oriented to the COVID-19 pandemic,through which we built a sentiment dictionary for the COVID-19 pandemic.The experiments based on the data annotated by psychologists indicated that the field sentiment dictionary generated through the sentiment dictionary extension framework effectively improved the accuracy,recall rate and F-value of sentiment classification.Lastly,we applied the constructed sentiment dictionary to the emotional polarity classification of Weibo comment data,obtained the trend curve for crowd stress emotions under the pandemic,analyzed and summarized the trends for crowd stress emotions during the COVID-19 outbreak in China,providing guidance for government decision-making and further analysis by psychologists.3.This paper proposes a parallel computing framework based on Hadoop Map Reduce EDEEV computing,namely HM-EDEEV,and establishes a corresponding Web service platform for psychological and emotional big data analysis of COVID-19(namely 2019 n Co VAS-PSA).The concrete contents include:First,a new distributed parallel computing framework based on Hadoop Map Reduce EDEEV algorithm,namely Hm-Edeev,is proposed.This framework has good acceleration ratio,scalability and scale growth.Secondly,Hadoop Streaming is introduced in hm-Edeev framework to solve the problem that EDEEV extension cannot cross languages.Therefore,the framework is still applicable to the affective dictionary extension based on algorithms in other industries and has certain universality.Secondly,we developed an easy-to-use crowd stress sentiment analysis service platform(2019n Co VAS-PSA)for crowds in the context of the COVID-19 pandemic based on the association analysis,which was combined with front-end technologies such as Echarts,Jquery,and H+ framework to provide a convenient and interactive data visualization tool for users.