Analyzing performance of high frequency currency rates prediction model using linear kernel SVR on historical data

Chanakya Serjam, Akito Sakurai

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

Abstract

We analyze the performance of various models constructed using linear kernel SVR and trained on historical bid data for high frequency currency trading. The bid tick data is converted into equally spaced (1 min) data. Different values for the number of training samples, number of features, and the length of the timeframes are used when conducting the experiments. These models are used to conduct simulated currency trading in the following year. We record the profits, hit ratios and number of trades executed from using these models. Our results indicate it is possible to obtain a profit as well as good hit ratio from a linear model trained only on historical data under certain pre-defined conditions. On examining the parameters for the linear models generated, we observe that a large number of models have all co-efficient values as negative while giving profit and good hit ratio, suggesting a simple yet effective trading strategy.

Original languageEnglish
Title of host publicationIntelligent Information and Database Systems - 9th Asian Conference, ACIIDS 2017, Proceedings
PublisherSpringer Verlag
Pages498-507
Number of pages10
Volume10191 LNAI
ISBN (Print)9783319544717
DOIs
Publication statusPublished - 2017
Event9th Asian Conference on Intelligent Information and Database Systems, ACIIDS 2017 - Kanazawa, Japan
Duration: 2017 Apr 32017 Apr 5

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10191 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other9th Asian Conference on Intelligent Information and Database Systems, ACIIDS 2017
CountryJapan
CityKanazawa
Period17/4/317/4/5

Fingerprint

Currency
Historical Data
Prediction Model
Hits
kernel
Profit
Linear Model
Profitability
Trading Strategies
Training Samples
Model
Coefficient
Experiment
Experiments

Keywords

  • Currency prediction
  • High frequency limit order book
  • Machine learning
  • Support vector regression (SVR)

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Serjam, C., & Sakurai, A. (2017). Analyzing performance of high frequency currency rates prediction model using linear kernel SVR on historical data. In Intelligent Information and Database Systems - 9th Asian Conference, ACIIDS 2017, Proceedings (Vol. 10191 LNAI, pp. 498-507). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10191 LNAI). Springer Verlag. https://doi.org/10.1007/978-3-319-54472-4_47

Analyzing performance of high frequency currency rates prediction model using linear kernel SVR on historical data. / Serjam, Chanakya; Sakurai, Akito.

Intelligent Information and Database Systems - 9th Asian Conference, ACIIDS 2017, Proceedings. Vol. 10191 LNAI Springer Verlag, 2017. p. 498-507 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10191 LNAI).

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

Serjam, C & Sakurai, A 2017, Analyzing performance of high frequency currency rates prediction model using linear kernel SVR on historical data. in Intelligent Information and Database Systems - 9th Asian Conference, ACIIDS 2017, Proceedings. vol. 10191 LNAI, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10191 LNAI, Springer Verlag, pp. 498-507, 9th Asian Conference on Intelligent Information and Database Systems, ACIIDS 2017, Kanazawa, Japan, 17/4/3. https://doi.org/10.1007/978-3-319-54472-4_47
Serjam C, Sakurai A. Analyzing performance of high frequency currency rates prediction model using linear kernel SVR on historical data. In Intelligent Information and Database Systems - 9th Asian Conference, ACIIDS 2017, Proceedings. Vol. 10191 LNAI. Springer Verlag. 2017. p. 498-507. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-54472-4_47
Serjam, Chanakya ; Sakurai, Akito. / Analyzing performance of high frequency currency rates prediction model using linear kernel SVR on historical data. Intelligent Information and Database Systems - 9th Asian Conference, ACIIDS 2017, Proceedings. Vol. 10191 LNAI Springer Verlag, 2017. pp. 498-507 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{8caa2b044f6a41909b3a6f91e58e9e4d,
title = "Analyzing performance of high frequency currency rates prediction model using linear kernel SVR on historical data",
abstract = "We analyze the performance of various models constructed using linear kernel SVR and trained on historical bid data for high frequency currency trading. The bid tick data is converted into equally spaced (1 min) data. Different values for the number of training samples, number of features, and the length of the timeframes are used when conducting the experiments. These models are used to conduct simulated currency trading in the following year. We record the profits, hit ratios and number of trades executed from using these models. Our results indicate it is possible to obtain a profit as well as good hit ratio from a linear model trained only on historical data under certain pre-defined conditions. On examining the parameters for the linear models generated, we observe that a large number of models have all co-efficient values as negative while giving profit and good hit ratio, suggesting a simple yet effective trading strategy.",
keywords = "Currency prediction, High frequency limit order book, Machine learning, Support vector regression (SVR)",
author = "Chanakya Serjam and Akito Sakurai",
year = "2017",
doi = "10.1007/978-3-319-54472-4_47",
language = "English",
isbn = "9783319544717",
volume = "10191 LNAI",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "498--507",
booktitle = "Intelligent Information and Database Systems - 9th Asian Conference, ACIIDS 2017, Proceedings",
address = "Germany",

}

TY - GEN

T1 - Analyzing performance of high frequency currency rates prediction model using linear kernel SVR on historical data

AU - Serjam, Chanakya

AU - Sakurai, Akito

PY - 2017

Y1 - 2017

N2 - We analyze the performance of various models constructed using linear kernel SVR and trained on historical bid data for high frequency currency trading. The bid tick data is converted into equally spaced (1 min) data. Different values for the number of training samples, number of features, and the length of the timeframes are used when conducting the experiments. These models are used to conduct simulated currency trading in the following year. We record the profits, hit ratios and number of trades executed from using these models. Our results indicate it is possible to obtain a profit as well as good hit ratio from a linear model trained only on historical data under certain pre-defined conditions. On examining the parameters for the linear models generated, we observe that a large number of models have all co-efficient values as negative while giving profit and good hit ratio, suggesting a simple yet effective trading strategy.

AB - We analyze the performance of various models constructed using linear kernel SVR and trained on historical bid data for high frequency currency trading. The bid tick data is converted into equally spaced (1 min) data. Different values for the number of training samples, number of features, and the length of the timeframes are used when conducting the experiments. These models are used to conduct simulated currency trading in the following year. We record the profits, hit ratios and number of trades executed from using these models. Our results indicate it is possible to obtain a profit as well as good hit ratio from a linear model trained only on historical data under certain pre-defined conditions. On examining the parameters for the linear models generated, we observe that a large number of models have all co-efficient values as negative while giving profit and good hit ratio, suggesting a simple yet effective trading strategy.

KW - Currency prediction

KW - High frequency limit order book

KW - Machine learning

KW - Support vector regression (SVR)

UR - http://www.scopus.com/inward/record.url?scp=85018595185&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85018595185&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-54472-4_47

DO - 10.1007/978-3-319-54472-4_47

M3 - Conference contribution

SN - 9783319544717

VL - 10191 LNAI

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 498

EP - 507

BT - Intelligent Information and Database Systems - 9th Asian Conference, ACIIDS 2017, Proceedings

PB - Springer Verlag

ER -