Sentiment analysis of comments on the Israel-Palestine conflict and showing geopolitical stance distribution

Abstract

Sentiment analysis has kept its landmark in Natural Language Processing by analyzing the text to extract sentimental value. The result usually goes with positive, negative, or neutral sentiment with necessary data preprocessing, data processing, and encoding. We collected the dataset from Kaggle which obtains the comments taken from Reddit posts regarding the Israel-Palestine conflict. The previous works and hypotheses were analyzedand the implementation of KNN and SVM is effective on those implementations. However, the whole concept of sentiment analysis is broadly focused on Natural Language Processing, and algorithms related to it should be used for actual accuracy and analysis. We applied Sentiment Intensity Analyzer and TextBlob algorithms and libraries to do the sentiment analysis of the desired dataset and compared them. These twoalgorithms have been widely described until today for their efficiency in sentimentanalysis. We found accuracy of 87.74% and 49.08% in the Sentiment Intensity Analyzer and TextBlob algorithm respectively. The best algorithm found here is the Sentiment Intensity Analyzer and we tested it accordingly. Finally, we showed Geopolitical Stance by applying manually entered input on topics, such as - Against Israel/Palestine, Supports Palestine, Neutral/Stance Not Clear. These two algorithms are easy and time-saving whereas traditional machine learning algorithm like KNN and SVM takes a lot of time and arealso not significant for sentiment analysis