Public opinion on microblogging sites, such as Twitter, is randomly distributed and so data mining such information offers many challenges technically. Writing in the International Journal of Autonomous and Adaptive Communications Systems, a team from China has now used a multi-visual clustering model to underpin a new algorithm to help them extract opinion from microblogging sites.
Lin-lin Li, Wei-zhen Hou, and Jing Liu of Renmin University of China in Beijing explain how microblogging often provides very timely and by virtue of its nature, succinct, public opinion data. Statistical analysis of such data might provide us with an almost real-time perspective on public opinion in various realms of activity whether political, commercial, artistic, scientific, or any other. Such opinion mining can help guide policy, marketing, and other areas of human endeavour so that it might jibe better with public opinion especially in areas of controversy.
The team has had much success but concedes that there is work still to do in removing invalid data prior to applying the algorithmic analysis. They also point out that there needs to be greater precision in the choice of experimental data so that the algorithm can be tuned to work more efficiently and efficaciously.
Li, L-l., Hou, W-z. and Liu, J. (2020) ‘Study on microblog public opinion data mining algorithm based on multi-visual clustering model‘, Int. J. Autonomous and Adaptive Communications Systems, Vol. 13, No. 2, pp.151-165.