TY - GEN
T1 - Stock Price Regression Based on Order Book Information
AU - Yoshida, Kenichi
AU - Sakurai, Akito
N1 - Funding Information:
This work was supported in part by JSPS KAKENHI Grant Number 25280114 and 25330266.
Publisher Copyright:
© 2016 IEEE.
PY - 2016/8/24
Y1 - 2016/8/24
N2 - Efficient market hypothesis, which entails the unpredictability of future stock prices, is widely accepted in financial market studies. In [1], we showed that a rule obtained by a simple analysis classifies short-term stock price changes with an 82.9% accuracy. In the analysis, the order book of high-frequency trading was the subject. The volume of high-frequency trading is increasing dramatically in these days, which is mainly responsible for short-term stock price changes, therefore, our study suggests the necessity of analyzing short-term market fluctuations caused by the high-frequency trading, an aspect that has not been well studied in conventional financial theories. In this paper, we extend our study to predict stock price by changing research framework from classification problem to regression problem. As expected based on [1], the regression model based on the proposed method can achieve very accurate results (e.g., correlation coefficient 0.48).
AB - Efficient market hypothesis, which entails the unpredictability of future stock prices, is widely accepted in financial market studies. In [1], we showed that a rule obtained by a simple analysis classifies short-term stock price changes with an 82.9% accuracy. In the analysis, the order book of high-frequency trading was the subject. The volume of high-frequency trading is increasing dramatically in these days, which is mainly responsible for short-term stock price changes, therefore, our study suggests the necessity of analyzing short-term market fluctuations caused by the high-frequency trading, an aspect that has not been well studied in conventional financial theories. In this paper, we extend our study to predict stock price by changing research framework from classification problem to regression problem. As expected based on [1], the regression model based on the proposed method can achieve very accurate results (e.g., correlation coefficient 0.48).
UR - http://www.scopus.com/inward/record.url?scp=84987978348&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84987978348&partnerID=8YFLogxK
U2 - 10.1109/COMPSAC.2016.199
DO - 10.1109/COMPSAC.2016.199
M3 - Conference contribution
AN - SCOPUS:84987978348
T3 - Proceedings - International Computer Software and Applications Conference
SP - 89
EP - 94
BT - Proceedings - 2016 IEEE 40th Annual Computer Software and Applications Conference Workshops, COMPSAC 2016
A2 - Liu, Ling
A2 - Milojicic, Dejan
A2 - Zhang, Zhiyong
A2 - Zhang, Zhiyong
A2 - Ahamed, Sheikh Iqbal
A2 - Sato, Hiroyuki
A2 - Cimato, Stevlio
A2 - Claycomb, William
A2 - Reisman, Sorel
A2 - Nakamura, Motonori
A2 - Lung, Chung Horng
A2 - Matskin, Mihhail
PB - IEEE Computer Society
T2 - 2016 IEEE 40th Annual Computer Software and Applications Conference, COMPSAC 2016
Y2 - 10 June 2016 through 14 June 2016
ER -