Forecasting of stock market values of airline companies by using social media data

The main objective of this study is to predictive the stock market value of Airline companies by using social media data and to investigate the investment feasibility based on this information. For this purpose, 236 764 tweets shared between 01.10.2019-06.02.2020 are collected from Twitter for Turkish Airlines (THY), Singapore Airlines (SIA), Lufthansa Airlines (DLH), Qantas Airlines (QFA), and Air France Airlines (AFR) companies by using Rapid Miner. In addition, daily stock market value data is obtained from investing.com website. The collected Twitter posts are classified into four sentiment categories: positive, negative, neutral, and none by making sentiment analysis with the MeaningCloud application. Then, the relationship between positive-negative tweet counts Net Promoter Score (NPS) and total tweet counts calculated with these emotion categories is studied, and the stock market value of the next day was investigated by using correlation analysis with SPSS statistics program. As a result of the correlation analysis, it was determined that there was a negative between the stock market value data and the number of citations on the Twitter platform for THY, DLH, and AFR companies, and a positive mid-rate correlation for SIA company. Finally, Auto-Model was used in the RapidMiner program to create a stock market value prediction model. "Gradient Boosted Trees" model has the lowest error prediction for THY and SIA companies. The increase or decrease of the stock market value of the next day is estimated correctly as 55,6% for THY and 57,1% for SIA. With this model, the stock market value of SIA on the next day is predicted with 5,1% and THY's with a 2,6% error rate using the number of commemorations. In the prediction model, it was observed that the NPS value had the highest weight-factor for SIA and the second-highest effect for THY after BIST-100 value. Consequently, it has been determined that the number of Twitter citations is a crucial data source that can be utilized in making investment decisions. Keywords: Social Media Analysis; Stock Market Value; Twitter; Correlation; Data Mining; Prediction Model.