Multiple kernel learning on time series data and social networks for stock price prediction

Shangkun Deng, Takashi Mitsubuchi, Kei Shioda, Tatsuro Shimada, Akito Sakurai

Research output: Chapter in Book/Report/Conference proceedingConference contribution

5 Citations (Scopus)

Abstract

This paper proposes a stock price prediction model, which extracts features from time series data, news, and comments on the news, for prediction of stock price and evaluates its performance. In this research, we do not take account of text contents of news and user comments, but just consider numerical features of news and communication dynamics appeared in comments on the Web as well as historical time series data. We model the stock price movements as a function of these input features and solve it as a regression problem in a Multiple Kernel Learning regression framework. Experimental results show that our proposed method consistently outperforms other baseline methods in terms of magnitude prediction measures such as MAE, MAPE and RMSE for three companies' stocks. They specifically show that the features other than stock prices themselves improved the performance.

Original languageEnglish
Title of host publicationProceedings - 10th International Conference on Machine Learning and Applications, ICMLA 2011
Pages228-234
Number of pages7
Volume2
DOIs
Publication statusPublished - 2011
Event10th International Conference on Machine Learning and Applications, ICMLA 2011 - Honolulu, HI, United States
Duration: 2011 Dec 182011 Dec 21

Other

Other10th International Conference on Machine Learning and Applications, ICMLA 2011
CountryUnited States
CityHonolulu, HI
Period11/12/1811/12/21

Fingerprint

Time series
Communication
Industry

Keywords

  • Communication Dynamics
  • Human Factors
  • Multiple Kernel Learning
  • Social Networks
  • Stock Price Prediction
  • Time Series Data

ASJC Scopus subject areas

  • Computer Science Applications
  • Human-Computer Interaction

Cite this

Deng, S., Mitsubuchi, T., Shioda, K., Shimada, T., & Sakurai, A. (2011). Multiple kernel learning on time series data and social networks for stock price prediction. In Proceedings - 10th International Conference on Machine Learning and Applications, ICMLA 2011 (Vol. 2, pp. 228-234). [6147679] https://doi.org/10.1109/ICMLA.2011.99

Multiple kernel learning on time series data and social networks for stock price prediction. / Deng, Shangkun; Mitsubuchi, Takashi; Shioda, Kei; Shimada, Tatsuro; Sakurai, Akito.

Proceedings - 10th International Conference on Machine Learning and Applications, ICMLA 2011. Vol. 2 2011. p. 228-234 6147679.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Deng, S, Mitsubuchi, T, Shioda, K, Shimada, T & Sakurai, A 2011, Multiple kernel learning on time series data and social networks for stock price prediction. in Proceedings - 10th International Conference on Machine Learning and Applications, ICMLA 2011. vol. 2, 6147679, pp. 228-234, 10th International Conference on Machine Learning and Applications, ICMLA 2011, Honolulu, HI, United States, 11/12/18. https://doi.org/10.1109/ICMLA.2011.99
Deng S, Mitsubuchi T, Shioda K, Shimada T, Sakurai A. Multiple kernel learning on time series data and social networks for stock price prediction. In Proceedings - 10th International Conference on Machine Learning and Applications, ICMLA 2011. Vol. 2. 2011. p. 228-234. 6147679 https://doi.org/10.1109/ICMLA.2011.99
Deng, Shangkun ; Mitsubuchi, Takashi ; Shioda, Kei ; Shimada, Tatsuro ; Sakurai, Akito. / Multiple kernel learning on time series data and social networks for stock price prediction. Proceedings - 10th International Conference on Machine Learning and Applications, ICMLA 2011. Vol. 2 2011. pp. 228-234
@inproceedings{55bbf079075c4fedace9bf2662187cfc,
title = "Multiple kernel learning on time series data and social networks for stock price prediction",
abstract = "This paper proposes a stock price prediction model, which extracts features from time series data, news, and comments on the news, for prediction of stock price and evaluates its performance. In this research, we do not take account of text contents of news and user comments, but just consider numerical features of news and communication dynamics appeared in comments on the Web as well as historical time series data. We model the stock price movements as a function of these input features and solve it as a regression problem in a Multiple Kernel Learning regression framework. Experimental results show that our proposed method consistently outperforms other baseline methods in terms of magnitude prediction measures such as MAE, MAPE and RMSE for three companies' stocks. They specifically show that the features other than stock prices themselves improved the performance.",
keywords = "Communication Dynamics, Human Factors, Multiple Kernel Learning, Social Networks, Stock Price Prediction, Time Series Data",
author = "Shangkun Deng and Takashi Mitsubuchi and Kei Shioda and Tatsuro Shimada and Akito Sakurai",
year = "2011",
doi = "10.1109/ICMLA.2011.99",
language = "English",
isbn = "9780769546070",
volume = "2",
pages = "228--234",
booktitle = "Proceedings - 10th International Conference on Machine Learning and Applications, ICMLA 2011",

}

TY - GEN

T1 - Multiple kernel learning on time series data and social networks for stock price prediction

AU - Deng, Shangkun

AU - Mitsubuchi, Takashi

AU - Shioda, Kei

AU - Shimada, Tatsuro

AU - Sakurai, Akito

PY - 2011

Y1 - 2011

N2 - This paper proposes a stock price prediction model, which extracts features from time series data, news, and comments on the news, for prediction of stock price and evaluates its performance. In this research, we do not take account of text contents of news and user comments, but just consider numerical features of news and communication dynamics appeared in comments on the Web as well as historical time series data. We model the stock price movements as a function of these input features and solve it as a regression problem in a Multiple Kernel Learning regression framework. Experimental results show that our proposed method consistently outperforms other baseline methods in terms of magnitude prediction measures such as MAE, MAPE and RMSE for three companies' stocks. They specifically show that the features other than stock prices themselves improved the performance.

AB - This paper proposes a stock price prediction model, which extracts features from time series data, news, and comments on the news, for prediction of stock price and evaluates its performance. In this research, we do not take account of text contents of news and user comments, but just consider numerical features of news and communication dynamics appeared in comments on the Web as well as historical time series data. We model the stock price movements as a function of these input features and solve it as a regression problem in a Multiple Kernel Learning regression framework. Experimental results show that our proposed method consistently outperforms other baseline methods in terms of magnitude prediction measures such as MAE, MAPE and RMSE for three companies' stocks. They specifically show that the features other than stock prices themselves improved the performance.

KW - Communication Dynamics

KW - Human Factors

KW - Multiple Kernel Learning

KW - Social Networks

KW - Stock Price Prediction

KW - Time Series Data

UR - http://www.scopus.com/inward/record.url?scp=84857836097&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84857836097&partnerID=8YFLogxK

U2 - 10.1109/ICMLA.2011.99

DO - 10.1109/ICMLA.2011.99

M3 - Conference contribution

SN - 9780769546070

VL - 2

SP - 228

EP - 234

BT - Proceedings - 10th International Conference on Machine Learning and Applications, ICMLA 2011

ER -