Prediction of Political and Local Conflicts in Bangladesh: An Event Analysis

• Published in: 2021 International Conference on Science & Contemporary Technologies (ICSCT) pp. 1 - 6
• Main Authors: Hasan, H M Mahmudul, Ahnaf, Adil, Hossain, Nahid
• Format: Conference Proceeding
• Language: English
• Published: IEEE 05.08.2021
• Subjects: Bangladesh; Conflict Analysis; Data models; Geology; Performance evaluation; Political and Local; Prediction; Prediction algorithms; Predictive models; Training; Urban areas
• DOI: 10.1109/ICSCT53883.2021.9642517

The international communities are trying to establish a comprehensive, precise, and valuable early warning system for conflict prevention involving political and local riots for many decades. The field of machine learning has some potential and promising components to develop this type of system. To predict and efficiently analyze all political conflicts in Bangladesh is the main motive of this paper. This study focuses on the time and geolocation of the conflict events, learning the event pattern, and predicting future conflicts in such areas. This study used the Naive Bayes algorithm to create a dataset blueprint that was later trained with Random Forrest to make a prediction model and used Tableau to map geolocations of predicting data for visualization. This study shows satisfactory results with 94.84 per cent accuracy. The dataset is available at https://tinyurl.com/hs9vxnu