Abstract—This paper presents a proposal to apply Latent Semantics Indexing to automatically classify Twitter tweets into different categories, in order to create a location-based geographic map of students’ civic engagement intensity and correlate it with social behavior. Since the work is at the proposal stage, the focus of this paper is on the proposed research methodology stemming from a pilot study we conducted with Facebook data. We implemented the methodology in a posting classification tool working with Facebook API. During our validation, the tool extracted 100 postings and classified them into five categories of politics, entertainment, science, technology, and daily life. Once adopted to analyzing tweets, we hope to contribute to the field by applying machine learning algorithms to the study of social behavior with focus on measuring youth civic engagement.
Index Terms—Classifier, latent semantics indexing, machine learning, pattern recognition.
S. Compion is with Interdisciplinary and Global Studies, Kean University, Union, NJ 07083 USA (e-mail: scompion@kean.edu).
J. J. Li is with School of Computer Science, Kean University, Union, NJ 07083 USA (e-mail: juli@kean.edu).
P. Croft, K. Ngoy and F. Qi are with the School of Sustainable and Environmental Sciences, Kean University, Union, NJ 07083 USA (e-mail: {pcroft|kngoy|fqi}@kean.edu).
Cite: S. Compion, P. Croft, J. J. Li, K. Ngoy, and F. Qi, "Tweet Semantic Classification in Civic Engagement Research," International Journal of Machine Learning and Computing vol. 8, no. 6, pp. 595-599, 2018.