Title: Early warning method of network public opinion communication crisis based on feature mining

Authors: Qing Cai

Addresses: School of Literature and Law, North China Institute of Science and Technology, Lang Fang 065201, China

Abstract: In order to overcome the problems of large P-R value error, high prediction error, and low recall rate in the process of online public opinion dissemination crisis warning, a feature mining-based online public opinion dissemination crisis warning method is proposed. Through MAP_reduce word segmentation processing technology, the word frequency characteristics of public opinion data are mined and processed. A time series model of public opinion was constructed, then data delays were eliminated. The signal characteristics of the crisis data transmitted by the sensor network public opinion through principal component analysis was obtained, and crisis warning was completed. The experimental results show that the precision value always maintains a high level with the change of the recall value, the slope of the curve is close to 45°, and the recall rate and precision rate are both greater than the traditional public opinion information early warning method.

Keywords: feature mining; network public opinion; communication crisis; early warning.

DOI: 10.1504/IJWBC.2022.125501

International Journal of Web Based Communities, 2022 Vol.18 No.3/4, pp.288 - 300

Received: 06 May 2021
Accepted: 05 Nov 2021
Published online: 12 Sep 2022 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article