|
|
Bimonthly Since 1986 |
ISSN 1004-9037
|
|
|
|
|
Publication Details |
Edited by: Editorial Board of Journal of Data Acquisition and Processing
P.O. Box 2704, Beijing 100190, P.R. China
Sponsored by: Institute of Computing Technology, CAS & China Computer Federation
Undertaken by: Institute of Computing Technology, CAS
Published by: SCIENCE PRESS, BEIJING, CHINA
Distributed by:
China: All Local Post Offices
|
|
|
|
|
|
|
|
|
|
Abstract
Based on the research into computational learning theory and pattern recognition, a new subject of computer science was born: machine learning. Continued usage of the Capital Asset Pricing Model (CAPM) based on information from the US equities markets allows for accurate price projections. Training on time series data for the whole stock universe and external macroeconomic factors allows the applicable Machine Learning models to significantly outperform the CAPM on out-of-sample (OOS) test data. Scores for categorization models varied widely; among the machine and deep learning models examined, the long short-term memory (LSTM) model showed greater accuracy. In most directional evaluation indicators, the experimental results show that traditional ML algorithms perform better. Furthermore, all ML algorithms are vulnerable to fluctuations in transaction cost, which can negatively affect trading performance. Yet, the effects of both explicit and implicit transaction costs on market activity are distinct. This research is important because it allows us to determine which algorithm is most lucrative across various markets.
Keyword
Recognition, Machine, Learning, Accuracy, Risk,
PDF Download (click here)
|
|
|
|
|