A Microblog Rumor Detection Model Integrating Multi-Dimensional Emotional Features

DOI: https://doi.org/10.62517/jbdc.202601204

Author(s)

Siqi Chen*

Affiliation(s)

School of Big Data and Artificial Intelligence, Jilin Business and Technology College, Changchun, Jilin, China *Corresponding Author

Abstract

To address the rapid spread and strong misleading nature of rumors on social media, this study proposes a microblog rumor detection method integrating textual features and multi-dimensional emotional features. Based on the Weibo21 public dataset, 3,200 text samples were randomly selected and manually annotated according to three emotional dimensions: panic, confrontation, and doubt. By comparing models including Logistic Regression, XGBoost, LightGBM, and a lightweight deep feature fusion model (repaired CNN-GRU-Attention model), the enhancement effect of emotional features on rumor detection performance was validated. Experimental results show that the Logistic Regression model combining textual and basic emotional features achieved the best performance among traditional models, with an F1-score of 0.8588, an improvement of 0.0057 compared to the text-only baseline. The lightweight deep feature fusion model further increased the F1-score to 0.8931 and accuracy to 0.8859, significantly outperforming traditional machine learning models. This study confirms that panic, confrontation, and doubt emotional features have significant correlations with rumor labels and can effectively enhance the model’s ability to identify rumor texts.

Keywords

Microblog Rumor Detection; Emotional Analysis; Multi-Feature Fusion; Deep Learning; Logistic Regression

References

[1] Liu, X.N., Hong, X.Y., Cao, Z.Y., Li, R.R., Wang, Z.S., Zhou, J.K., Lu, H.Y. (2025) Social Media Rumor Detection: Methods, Challenges and Trends. Computer Engineering and Applications, 61(11): 31-50. [2] Meng, J.N., Wang, X.P., Li, T., Liu, S., Zhao, D. (2022) Cross-modal Rumor Detection Based on Adversarial Neural Network. Data Analysis and Knowledge Discovery, 6(12): 32-42. [3] Feng, R.J., Zhang, H.J., Pan, W.M. (2023) Early Microblog Rumor Detection Based on Pre-trained Language Model. Computer and Digital Engineering, 51(05): 1075-1080+1184. [4] Liu, X.J., Xu, Y.C., Dong, F.M. (2020) Microblog Rumor Detection Combining Text and User Profile Data. Information and Communications, (12): 39-43. [5] Ma, M.F., Chen, J.H., Li, Y., Zhang, C. (2025) MFLAN: A Multi-feature Fusion Rumor Detection Model Based on Improved GAT. Computer Engineering, 51(08): 181-189. [6] Liu, T.T., Zhu, W.D., Liu, G.Y. (2018) Research Progress in Text Classification Based on Deep Learning. Electric Power Information and Communication Technology, 16(03): 1-7. [7] Li, Y., Dong, H.B. (2018) Text Sentiment Analysis Based on Feature Fusion of CNN and BiLSTM Network. Computer Applications, 38(11): 3075-3080. [8] Feng, R.J., Zhang, H.J., Pan, W.M. (2021) Microblog Rumor Detection Based on Sentiment Analysis and Transformer Model. Computer and Modernization, (10): 1-7. [9] Yang, Y.J., Wang, L., Wang, Y.H. (2021) Rumor Detection Research Integrating Source Information and Gated Graph Neural Network. Journal of Computer Research and Development, 58(07): 1412-1424. [10] Wang, L.Y., Liu, C.H., Cai, D.B., Lu, T. (2019) Chinese Text Sentiment Analysis by Introducing Attention Mechanism in CNN-BiGRU Network. Computer Applications, 39(10): 2841-2846. [11] Wang, F., Guo, J.J., Yu, Z.T. (2022) Multi-task Joint Rumor Detection Method Integrating Comments. Computer Engineering and Science, 44(09): 1702-1710. [12] Jiang, B.Y., Dan, Z.P., Dong, F.M., Zhang, H.Z., Liu, Z.Y. (2023) Multimodal Rumor Detection Based on Dual Pre-trained Transformer and Cross Attention. Foreign Electronic Measurement Technology, 42(04): 149-157. DOI:10.19652/j.cnki.femt.2304724 [13] Gao, Z., Dan, Z.P., Dong, F.M., Zhang, Y.K., Zhang, H.Z. (2024) Social Media Rumor Detection Based on Multi-level Unauthenticity Propagation Structure. Journal of Chinese Information Processing, 38(02): 142-154. [14] Zhao, R.M., Xiong, X., Ju, S.G., Li, Z.Z., Xie, C. (2020) Chinese Implicit Sentiment Analysis Based on Hybrid Neural Network. Journal of Sichuan University (Natural Science Edition), 57(02): 264-270. [15] Cheng, Y., Sun, H., Chen, H.M., Li, M., Cai, Y.Y., Cai, Z. (2021) Text Sentiment Analysis Capsule Model Integrating Convolutional Neural Network and Bidirectional GRU. Journal of Chinese Information Processing, 35(05): 118-129. [16] Su, X., Yu, K., Wu, X.F. (2023) Rumor Detection Method Based on Hierarchical Gated Interactive Fusion Network. Journal of Beijing University of Posts and Telecommunications, 46(04): 97-102. [17] Qiang, Z.S., Gu, Y.J. (2023) Social Media Rumor Detection Model Based on Multi-modal Heterogeneous Graph. Data Analysis and Knowledge Discovery, 7(11): 68-78. [18]Duan, D.G., Bai, C.Y., Han, Z.M., Xiong, H.T. (2022) Social Media Rumor Detection Method Based on Multi-pass Influence. Computer Engineering, 48(10): 138-145+157. [19] Wen, T.X., Gao, Q. (2023) BiLSTM-CNN-ECA Rumor Detection Model Combining Multi-granularity Semantics Inside and Outside Microblogs. Information Research, (05): 78-84. [20] Li, A., Dan, Z.P., Dong, F.M., Liu, L.W., Feng, Y. (2020) Rumor Detection Method Based on Improved Generative Adversarial Network. Journal of Chinese Information Processing, 34(09): 78-88. [21] Jin, Z.G., Yang, Y. (2019) Semi-supervised Sentiment Analysis Model Based on User Correlation. Journal of Harbin Institute of Technology, 51(05): 50-56.