Article: Spatio-temporal convolutional networks empowering political ideology trend prediction on social media platforms Journal: International Journal of Information and Communication Technology (IJICT) 2025 Vol.26 No.44 pp.39 - 57 Abstract: This paper proposes a prediction model based on a spatio-temporal graph convolutional network to capture the spatio-temporal dependencies of user interactions and improve the accuracy of predicting trends in the dissemination of ideological and political themes. Experimental results show that this method achieves an average improvement of 6.8% in the F1 score and a reduction of 7.2% in the prediction root mean square error for key metrics such as the probability of ideological and political hot topics appearing in the coming week and changes in regional sentiment distribution. This model effectively integrates the spatial structural information of social networks with dynamic temporal features, providing a reliable computational tool for quantitative analysis and forward-looking assessment of ideological and political dynamics in social media environments. These aid relevant departments in timely sensing and guiding the online ideological and political ecosystem. Inderscience Publishers - linking academia, business and industry through research

Title: Spatio-temporal convolutional networks empowering political ideology trend prediction on social media platforms

Authors: Wencheng Liu; Xiaojuan Deng

Addresses: School of Marxism, Hunan Open University, Changsha, 410004, China ' School of Economics and Management, Shaoyang University, Shaoyang, 422000, China

Abstract: This paper proposes a prediction model based on a spatio-temporal graph convolutional network to capture the spatio-temporal dependencies of user interactions and improve the accuracy of predicting trends in the dissemination of ideological and political themes. Experimental results show that this method achieves an average improvement of 6.8% in the F1 score and a reduction of 7.2% in the prediction root mean square error for key metrics such as the probability of ideological and political hot topics appearing in the coming week and changes in regional sentiment distribution. This model effectively integrates the spatial structural information of social networks with dynamic temporal features, providing a reliable computational tool for quantitative analysis and forward-looking assessment of ideological and political dynamics in social media environments. These aid relevant departments in timely sensing and guiding the online ideological and political ecosystem.

Keywords: spatio-temporal convolutional networks; social media platforms; ideological and political communication.

DOI: 10.1504/IJICT.2025.150407

International Journal of Information and Communication Technology, 2025 Vol.26 No.44, pp.39 - 57

Received: 08 Sep 2025
Accepted: 20 Oct 2025
Published online: 12 Dec 2025 *

Title: Spatio-temporal convolutional networks empowering political ideology trend prediction on social media platforms

Keep up-to-date