Text Mining of Twitter Data for Public Sentiment Analysis of Staple Foods Price Changes
Isti Surjandari1, Muthia Szami Naffisah1,2, and M. Irfan Prawiradinata1,2
1.Department of Industrial Engineering Faculty of Engineering, Universitas Indonesia
2.Department of Economics, Universitas Indonesia
2.Department of Economics, Universitas Indonesia
Abstract—Millions of users share their opinions on twitter, making it valuable platform for tracking and analyzing public opinion. Such analysis can provide critical information for decision maker in various domains. In this study, we examine public sentiment analysis of staple foods price changes in Indonesia based on twitter data. Text mining was used for classifying tweets into positive and negative sentiment. Then association between the type of staple foods and sentiment classes were analyzed using Chi Square test and Marascuillo procedure. Results show that Support Vector Machine (SVM) classifier produce higher accuracy than Naïve Bayes and Decision Trees. Also, the price of milk, eggs and red onion had the most significant association to the negative sentiment compared to other commodities.
Index Terms—text mining, sentiment analysis, marascuilo, food prices, twitter
Cite: Isti Surjandari, Muthia Szami Naffisah, and M. Irfan Prawiradinata, "Text Mining of Twitter Data for Public Sentiment Analysis of Staple Foods Price Changes," Journal of Industrial and Intelligent Information, Vol. 3, No. 3, pp. 253-257, September 2015. doi: 10.12720/jiii.3.3.253-257
Index Terms—text mining, sentiment analysis, marascuilo, food prices, twitter
Cite: Isti Surjandari, Muthia Szami Naffisah, and M. Irfan Prawiradinata, "Text Mining of Twitter Data for Public Sentiment Analysis of Staple Foods Price Changes," Journal of Industrial and Intelligent Information, Vol. 3, No. 3, pp. 253-257, September 2015. doi: 10.12720/jiii.3.3.253-257