Activism via attention: interpretable spatiotemporal learning to forecast protest activities
Abstract The diffusion of new information and communication technologies—social media in particular—has played a key role in social and political activism in recent decades. In this paper, we propose a theory-motivated, spatiotemporal learning approach, ActAttn, that leverages social movement theori...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
SpringerOpen
2019-02-01
|
| Series: | EPJ Data Science |
| Subjects: | |
| Online Access: | http://link.springer.com/article/10.1140/epjds/s13688-019-0183-y |
| _version_ | 1811324925950033920 |
|---|---|
| author | Ali Mert Ertugrul Yu-Ru Lin Wen-Ting Chung Muheng Yan Ang Li |
| author_facet | Ali Mert Ertugrul Yu-Ru Lin Wen-Ting Chung Muheng Yan Ang Li |
| author_sort | Ali Mert Ertugrul |
| collection | DOAJ |
| description | Abstract The diffusion of new information and communication technologies—social media in particular—has played a key role in social and political activism in recent decades. In this paper, we propose a theory-motivated, spatiotemporal learning approach, ActAttn, that leverages social movement theories and a deep learning framework to examine the relationship between protest events and their social and geographical contexts as reflected in social media discussions. To do so, we introduce a novel predictive framework that incorporates a new design of attentional networks, and which effectively learns the spatiotemporal structure of features. Our approach is not only capable of forecasting the occurrence of future protests, but also provides theory-relevant interpretations—it allows for interpreting what features, from which places, have significant contributions on the protest forecasting model, as well as how they make those contributions. Our experiment results from three movement events indicate that ActAttn achieves superior forecasting performance, with interesting comparisons across the three events that provide insights into these recent movements. |
| first_indexed | 2024-04-13T14:23:56Z |
| format | Article |
| id | doaj.art-4cb80f7257aa41bb85c21d82334df51d |
| institution | Directory Open Access Journal |
| issn | 2193-1127 |
| language | English |
| last_indexed | 2024-04-13T14:23:56Z |
| publishDate | 2019-02-01 |
| publisher | SpringerOpen |
| record_format | Article |
| series | EPJ Data Science |
| spelling | doaj.art-4cb80f7257aa41bb85c21d82334df51d2022-12-22T02:43:22ZengSpringerOpenEPJ Data Science2193-11272019-02-018112610.1140/epjds/s13688-019-0183-yActivism via attention: interpretable spatiotemporal learning to forecast protest activitiesAli Mert Ertugrul0Yu-Ru Lin1Wen-Ting Chung2Muheng Yan3Ang Li4School of Computing and Information, University of PittsburghSchool of Computing and Information, University of PittsburghDepartment of Psychology in Education, School of Education, University of PittsburghSchool of Computing and Information, University of PittsburghSchool of Computing and Information, University of PittsburghAbstract The diffusion of new information and communication technologies—social media in particular—has played a key role in social and political activism in recent decades. In this paper, we propose a theory-motivated, spatiotemporal learning approach, ActAttn, that leverages social movement theories and a deep learning framework to examine the relationship between protest events and their social and geographical contexts as reflected in social media discussions. To do so, we introduce a novel predictive framework that incorporates a new design of attentional networks, and which effectively learns the spatiotemporal structure of features. Our approach is not only capable of forecasting the occurrence of future protests, but also provides theory-relevant interpretations—it allows for interpreting what features, from which places, have significant contributions on the protest forecasting model, as well as how they make those contributions. Our experiment results from three movement events indicate that ActAttn achieves superior forecasting performance, with interesting comparisons across the three events that provide insights into these recent movements.http://link.springer.com/article/10.1140/epjds/s13688-019-0183-yInterpretable spatiotemporal learningEvent forecastingCivil unrestProtest activities |
| spellingShingle | Ali Mert Ertugrul Yu-Ru Lin Wen-Ting Chung Muheng Yan Ang Li Activism via attention: interpretable spatiotemporal learning to forecast protest activities EPJ Data Science Interpretable spatiotemporal learning Event forecasting Civil unrest Protest activities |
| title | Activism via attention: interpretable spatiotemporal learning to forecast protest activities |
| title_full | Activism via attention: interpretable spatiotemporal learning to forecast protest activities |
| title_fullStr | Activism via attention: interpretable spatiotemporal learning to forecast protest activities |
| title_full_unstemmed | Activism via attention: interpretable spatiotemporal learning to forecast protest activities |
| title_short | Activism via attention: interpretable spatiotemporal learning to forecast protest activities |
| title_sort | activism via attention interpretable spatiotemporal learning to forecast protest activities |
| topic | Interpretable spatiotemporal learning Event forecasting Civil unrest Protest activities |
| url | http://link.springer.com/article/10.1140/epjds/s13688-019-0183-y |
| work_keys_str_mv | AT alimertertugrul activismviaattentioninterpretablespatiotemporallearningtoforecastprotestactivities AT yurulin activismviaattentioninterpretablespatiotemporallearningtoforecastprotestactivities AT wentingchung activismviaattentioninterpretablespatiotemporallearningtoforecastprotestactivities AT muhengyan activismviaattentioninterpretablespatiotemporallearningtoforecastprotestactivities AT angli activismviaattentioninterpretablespatiotemporallearningtoforecastprotestactivities |