A Deep Learning Model for Detecting and Classifying Protest Demand from Social Media

• In Event: Deep Learning Applications in Political Science

Brief Overview

We introduce a novel model that combines scene text recognition and text classification models to detect and classify protest demands. Compared to past research, it can not only better detect the rise of protest but can classify types of protests.

Authors

• Xuan Luo, Tokyo Foundation for Policy Research

• Sota Kato, International University of Japan

• Budrul Ahsan, Philips Japan

• Takafumi Nakanishi, Musashino University