Hybrid Method of Multiple Kernel Learning and Genetic Algorithm for Forecasting Short-Term Foreign Exchange Rates

Shangkun Deng, Kazuki Yoshiyama, Takashi Mitsubuchi, Akito Sakurai

Research output: Contribution to journalArticle

26 Citations (Scopus)

Abstract

Our proposed prediction and learning method is a hybrid referred to as MKL-GA, which combines multiple kernel learning (MKL) for regression (MKR) and a genetic algorithm (GA) to construct the trading rules. In this study, we demonstrate that the evaluation criteria used to examine the effectiveness of a financial market price forecasting method should be the profit and profit-risk ratio, rather than errors in prediction. Thus, it is necessary to use a price prediction method and a trading rules learning method. We tested the proposed method on the foreign exchange market for the USD/JPY currency pair, where the features used for prediction were extracted from the trading history of the three main currency pairs with three different short-term horizons. MKR is essential for utilizing the information contained in many of the features derived from different information sources and for various representations of the same information source. The GA is essential for generating trading rules, which are described using a mixture of discrete structures and continuous parameters. First, the MKR predicts the change in the exchange rate based on technical indicators such as the moving average convergence and divergence of the three currency pairs. Next, the GA generates a trading rule by combining the results of the MKR with several commonly used overbought/oversold technical indicators. The experimental results show that the proposed hybrid method outperforms other baseline methods in terms of the returns and return-risk ratio. In addition, the kernel weights employed for different currency pairs and the different time horizons used in the MKR step, as well as the trading strategy generated in the GA step, should be beneficial during actual trading.

Original languageEnglish
Pages (from-to)1-41
Number of pages41
JournalComputational Economics
DOIs
Publication statusAccepted/In press - 2013

Fingerprint

Learning algorithms
Genetic algorithms
Profitability
History
Kernel
Foreign exchange rates
Hybrid method
Learning algorithm
Genetic algorithm
Trading rules
Prediction
Currency
Financial markets
Information sources
Profit
Learning methods

Keywords

  • Financial prediction
  • FX trading
  • Genetic algorithm
  • MKL-GA hybrid method
  • Multiple kernel learning
  • Technical indicators

ASJC Scopus subject areas

  • Economics, Econometrics and Finance (miscellaneous)
  • Computer Science Applications

Cite this

Hybrid Method of Multiple Kernel Learning and Genetic Algorithm for Forecasting Short-Term Foreign Exchange Rates. / Deng, Shangkun; Yoshiyama, Kazuki; Mitsubuchi, Takashi; Sakurai, Akito.

In: Computational Economics, 2013, p. 1-41.

Research output: Contribution to journalArticle

Deng, Shangkun ; Yoshiyama, Kazuki ; Mitsubuchi, Takashi ; Sakurai, Akito. / Hybrid Method of Multiple Kernel Learning and Genetic Algorithm for Forecasting Short-Term Foreign Exchange Rates. In: Computational Economics. 2013 ; pp. 1-41.
@article{ca82301cd42a48c29e4e27333a5b370c,
title = "Hybrid Method of Multiple Kernel Learning and Genetic Algorithm for Forecasting Short-Term Foreign Exchange Rates",
abstract = "Our proposed prediction and learning method is a hybrid referred to as MKL-GA, which combines multiple kernel learning (MKL) for regression (MKR) and a genetic algorithm (GA) to construct the trading rules. In this study, we demonstrate that the evaluation criteria used to examine the effectiveness of a financial market price forecasting method should be the profit and profit-risk ratio, rather than errors in prediction. Thus, it is necessary to use a price prediction method and a trading rules learning method. We tested the proposed method on the foreign exchange market for the USD/JPY currency pair, where the features used for prediction were extracted from the trading history of the three main currency pairs with three different short-term horizons. MKR is essential for utilizing the information contained in many of the features derived from different information sources and for various representations of the same information source. The GA is essential for generating trading rules, which are described using a mixture of discrete structures and continuous parameters. First, the MKR predicts the change in the exchange rate based on technical indicators such as the moving average convergence and divergence of the three currency pairs. Next, the GA generates a trading rule by combining the results of the MKR with several commonly used overbought/oversold technical indicators. The experimental results show that the proposed hybrid method outperforms other baseline methods in terms of the returns and return-risk ratio. In addition, the kernel weights employed for different currency pairs and the different time horizons used in the MKR step, as well as the trading strategy generated in the GA step, should be beneficial during actual trading.",
keywords = "Financial prediction, FX trading, Genetic algorithm, MKL-GA hybrid method, Multiple kernel learning, Technical indicators",
author = "Shangkun Deng and Kazuki Yoshiyama and Takashi Mitsubuchi and Akito Sakurai",
year = "2013",
doi = "10.1007/s10614-013-9407-6",
language = "English",
pages = "1--41",
journal = "Computer Science in Economics and Management",
issn = "0921-2736",
publisher = "Springer Netherlands",

}

TY - JOUR

T1 - Hybrid Method of Multiple Kernel Learning and Genetic Algorithm for Forecasting Short-Term Foreign Exchange Rates

AU - Deng, Shangkun

AU - Yoshiyama, Kazuki

AU - Mitsubuchi, Takashi

AU - Sakurai, Akito

PY - 2013

Y1 - 2013

N2 - Our proposed prediction and learning method is a hybrid referred to as MKL-GA, which combines multiple kernel learning (MKL) for regression (MKR) and a genetic algorithm (GA) to construct the trading rules. In this study, we demonstrate that the evaluation criteria used to examine the effectiveness of a financial market price forecasting method should be the profit and profit-risk ratio, rather than errors in prediction. Thus, it is necessary to use a price prediction method and a trading rules learning method. We tested the proposed method on the foreign exchange market for the USD/JPY currency pair, where the features used for prediction were extracted from the trading history of the three main currency pairs with three different short-term horizons. MKR is essential for utilizing the information contained in many of the features derived from different information sources and for various representations of the same information source. The GA is essential for generating trading rules, which are described using a mixture of discrete structures and continuous parameters. First, the MKR predicts the change in the exchange rate based on technical indicators such as the moving average convergence and divergence of the three currency pairs. Next, the GA generates a trading rule by combining the results of the MKR with several commonly used overbought/oversold technical indicators. The experimental results show that the proposed hybrid method outperforms other baseline methods in terms of the returns and return-risk ratio. In addition, the kernel weights employed for different currency pairs and the different time horizons used in the MKR step, as well as the trading strategy generated in the GA step, should be beneficial during actual trading.

AB - Our proposed prediction and learning method is a hybrid referred to as MKL-GA, which combines multiple kernel learning (MKL) for regression (MKR) and a genetic algorithm (GA) to construct the trading rules. In this study, we demonstrate that the evaluation criteria used to examine the effectiveness of a financial market price forecasting method should be the profit and profit-risk ratio, rather than errors in prediction. Thus, it is necessary to use a price prediction method and a trading rules learning method. We tested the proposed method on the foreign exchange market for the USD/JPY currency pair, where the features used for prediction were extracted from the trading history of the three main currency pairs with three different short-term horizons. MKR is essential for utilizing the information contained in many of the features derived from different information sources and for various representations of the same information source. The GA is essential for generating trading rules, which are described using a mixture of discrete structures and continuous parameters. First, the MKR predicts the change in the exchange rate based on technical indicators such as the moving average convergence and divergence of the three currency pairs. Next, the GA generates a trading rule by combining the results of the MKR with several commonly used overbought/oversold technical indicators. The experimental results show that the proposed hybrid method outperforms other baseline methods in terms of the returns and return-risk ratio. In addition, the kernel weights employed for different currency pairs and the different time horizons used in the MKR step, as well as the trading strategy generated in the GA step, should be beneficial during actual trading.

KW - Financial prediction

KW - FX trading

KW - Genetic algorithm

KW - MKL-GA hybrid method

KW - Multiple kernel learning

KW - Technical indicators

UR - http://www.scopus.com/inward/record.url?scp=84887581269&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84887581269&partnerID=8YFLogxK

U2 - 10.1007/s10614-013-9407-6

DO - 10.1007/s10614-013-9407-6

M3 - Article

AN - SCOPUS:84887581269

SP - 1

EP - 41

JO - Computer Science in Economics and Management

JF - Computer Science in Economics and Management

SN - 0921-2736

ER -