Combining Google Trends for the Prediction of Monthly Revenue of Firms in Taiwan Automobile Industry

• Main Authors: Chu-Yuan Ni, 朱元妮
• Other Authors: Chih-Cheng Chen
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/2pkd6e

碩士 === 元智大學 === 資訊管理學系 === 107 === The aim of this study is to investigate that whether the data of Google Trends can enhance the prediction for the monthly revenue of firms in Taiwan automobile industry. The top three companies in Taiwan's automobile industry are selected as the research objects. In addition to the revenue, we also utilize their searching volumes in Google Trends with their company’s name, stock codes and their brand names of their cars are chosen as the keywords. In terms of the research method, we adopt time series models, the VAR models and simple linear regression models. Then we conduct the error analysis on the prediction results to choose the most fitted model. The result validates that the VAR model that combines Google Trends and revenue data is the most fitted model for the revenue prediction of these three firms as compared with other models. In other words, which using Google Trends can improve the prediction of revenues of firms in Taiwan’s automobile industry. We believe our method can be generalized and to be applied for other companies and industries, and this method is also good foe the stock investors because it offers better prediction of revenue for the listed companies.