Mechanisms of data mining in analyzing the effects and trends of news dissemination

Abstract

Abstract Data mining techniques can help news organizations and media practitioners extract valuable information from a large amount of unorganized information, such as by analyzing the effect of news dissemination. In data mining, multiple linear regression is a widely used technical model. In this paper, we utilize the classic “cognition-attitude-behavior” analysis framework to construct a regression analysis model of the news communication effect. We solve the positional parameters according to the principle of least squares, ensuring that the regression value and all observations have the minimum residual square. We then solve the system of linear equations using Clem’s law and the Gaussian elimination method. The constructed models are used to analyze the news dissemination effect of Guangzhou’s city image on traditional news platforms and microblog self-media platforms, respectively. It is found that the news videos of “cultural image” and “ecological image” have a significant positive effect on the news dissemination effect. The longer the duration of the news video, the better the dissemination effect. In terms of the influence of event-related online opinion leaders on news dissemination, the number of comments and likes have a significant positive influence, with coefficients of 0.778 and 0.059, respectively, and both are significant at the 1% level. In addition to the number of comments and likes, the influencing factors on the dissemination of related news are the number of fans, whether there is V authentication, and whether there is microblogging membership. After conducting two empirical data mining analyses, we verified the validity of the multiple linear regression model in this paper by analyzing the effect and trend of news dissemination.