Sentiment based information retrieval framework for cultural psychology

Abstract

People share their opinion and information through social networks platforms such as Twitter, You Tube, and Facebook. Their shared opinions towards certain issues are sentiments that could be productive, constructive, or possibly controversial. These opinions are positive or negative sentiments. Sentiment analysis are done mainly on marketing and political issues. They focused on trends to improve their services to achieve their targeted audiences and customers. However, there is yet a need to conduct sentiment analysis on cultural psychology issues. Thus, this research aimed to analyse and categorize the sentiments people shared on a social network pertaining to the selected issues on topics in cultural psychology. The Zheng and Fang model was adapted for sentiment analysis. Three social networks were selected; You Tube, Facebook, and Twitter that offer search capability enabling the retrieval of posted comments and opinions. A sample of 100 cases based on the selected topics have been collected and formulated as queries. The queries retrieved the sentiments. The identified sentiments were analysed and classified as positive and negative and topically categorized based on a value system using WordStat8 and LightSIDE toolkit. The Prabowo and Thelwall combined model of sentiment analysis was referred to for categorization. The outcomes included a pool of positive and negative sentiments; and topic categorization developed based on sentiment analysis. Kappa, recall, precision and F-Scores were reported to range from -0.01 to 0.23, 0.06 to 1.00, 0.14 to 0.96, and 0.04 to 0.86 correspondingly. Overall, Kappa, precision, and F-scores ranged from very low to high ratios, except for the perfect recall.