Applying Sentiment Analysis Techniques in Social Media Data About Threat of Armed Conflicts Using Two Times Series Models

Abstract

The growing cases of armed conflicts over the past couple of decades have dramatically affected social landscapes and people's lives across the globe, urging everyone to find ways to minimize the negative consequences of the conflicts. Social media provides an inexhaustible data source that can be used in understanding the evolution of such conflicts. This chapter focuses on Syria-USA and Iran-USA relations to presents an approach to armed conflict analysis and examines the Russia-Ukraine conflicts by performing sentiment analysis on the text dataset as well as on a vocabulary data. All conflicts generate a social media news threat time series (TTS) that is used as input to the P-model algorithm to generate the endogenous time series. The following uses the TTS and endogenous time series for both conflicts as input to the deep-learning-LSTM neural network. Finally, this chapter compares the prediction result of the Russia-Ukraine TTS analysis with the Russia-Ukraine endogenous series using the P-model algorithm.