Affiliations: [a]
College of Computer Science and Engineering, Shandong University of Science and Technology, Jinan, China
| [b]
Department of Management Information Systems, National Chung Hsing University, Taichung City, Taiwan
| [c]
Department of Information and Finance Management, National Taipei University of Technology, Taipei, Taiwan
Correspondence:
[*]
Corresponding author. Meng-Hsiun Tsai, Department of Management Information Systems, National Chung Hsing University, Taichung City, Taiwan. E-mail: mht@nchu.edu.tw.
Abstract: With the rise in popularity of personal computers and decreasing cost, even a personal computer can execute complex and large calculations. So more researchers can invest in AI and machine learning. Humans can’t handle massive data sets or data that requires a long time to read and evaluate, whereas big data frameworks can read and analyze in a reasonable time. So relevant research has increased recently. In the social sciences, machine learning is used to forecast future trends and the index trend. Keeping up with current events is crucial nowadays to debate countermeasures in time. This study combines economic indicators from 1988 to 2017 with leading indicators and other types of indicators. The recurrent neural network model predicts economic index trends and tests multiple variables. The proposed methods measure the error in predicting future trends in different models to learn which indicators work well together.
Keywords: Machine learning, leading indicator, recurrent neural network, long short term memory, forecast trend
