Affiliations: [a] Business School, Lingnan Normal University, Zhanjiang, Guangdong 524048, China | [b] Guangdong Coastal Economic Belt Development Research Center, Zhanjiang, Guangdong 524088, China | E-mail: zufengzzz@163.com
Correspondence:
[*]
Guangdong Coastal Economic Belt Development Research Center, Zhanjiang, Guangdong 524088, China.
E-mail: zufengzzz@163.com.
Abstract: This study aims to gain insights into the basic information and behavioral characteristics of the drug abusers and provide references for drug prevention, control, and correctional strategies. First, the python development tool was used to crawl 8494 posts from 1725 users in the forum of “Dynamic Control Bar” in the Baidu Tieba. The data were cleaned and organized. Subsequently, the content of the posts in text was analyzed using a mixture of topic model, sentiment analysis, and relevance analysis. The result of the LDA indicated that the drug abusers were concerned about the living conditions of this population in their home communities, regular checkups and management by government staff, perceived social discrimination and inconvenience of living in a restrained environment, problems they encountered when consulting with each other in terms of regular medical checkups, recollection of how they came to use drugs, as well as emotions of regret. The result of the emotional analysis indicated that this population was emotionally disturbed and had more negative emotional values, but the above values were stable. Internet information dissemination is of great significance to public opinion dissemination that can indicate the real opinions and attitudes of all social strata to drug abusers, especially the discrimination, stigmatization, and labelling of drug abusers by the general public. Disseminating content to drug abusers about their problems can help them start a new life. Furthermore, the government should guide the attitudes and emotions of this population to help them start a new, more positive life.
Keywords: Drug abuse, dynamic control, latent dirichlet allocation, sentiment analysis, internet public opinion
