SENTIMENT ANALYSIS AND ITS EFFECT OF INVESTOR’S PERCEPTION FOR ANTICIPATING EQUITY VALUATION UPDATE UNDER STOCK RIGHT-ISSUE POLICY

• Main Author: Salim, Ervin
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/78576
• Institution: Institut Teknologi Bandung

Investor and trader sometimes easily affected by the others in trading stocks, especially when they see the news, rumors,also the response by the public, and at the end can influence how investors taking decisions. All this information can be seen in online media such as online news and online forum. This study will discuss those effects under special circumstance: new right issues activity. Right issue is an offer to existing shareholders to purchase additional new shares of the company. Using Big Data Analytics approach, this study tries to measure sentiment and perceptions of investor, combined with the standard valuation model that use historical transaction data and specific firm fundamental. Earlier intensive research conduct by Deddy Koesrindartoto, et.al in 2015, shows that not all source of information can be used to measure sentiments and perceptions accurately. Therefore, this study will select Stockbit® (www.stockbit.com) as the main source of information in measuring investors’ sentiment and perception. The unstructured data (such as statements, analysis, and other comments) related to rightissue stocks event inside the Stockbit® social media platform are being crawled and collected. These data processed using basket (bag) of words and at the end, each of statement results in aquantitative measure of sentiment score. These data will be processed together with the more traditional quantitative data such as historical transaction data and fundamental firm value. The methodology of analysis is employing the decision tree process and neural networks prediction model. In this study, we use a combination of script software such as PHP and research graded prediction tools software such as RapidMiner®. The results of this study show that sentiment or perception is becoming important factors in predicting the future value of right-issue stocks. In fact, it gave stronger relations and more significant effects compared to the traditional factors such as historical transaction data and fundamental firm data. These results show that the Big Data Analytic approach especially in measuring sentiment and perception has potential power for being used as among the future tools for investors.