An ensemble of LSTM neural networks for high-frequency stock market classification

Svetlana Borovkova, Ioannis Tsiamas

Research output: Contribution to JournalArticleAcademicpeer-review

Abstract

We propose an ensemble of long–short-term memory (LSTM) neural networks for intraday stock predictions, using a large variety of technical analysis indicators as network inputs. The proposed ensemble operates in an online way, weighting the individual models proportionally to their recent performance, which allows us to deal with possible nonstationarities in an innovative way. The performance of the models is measured by area under the curve of the receiver operating characteristic. We evaluate the predictive power of our model on several US large-cap stocks and benchmark it against lasso and ridge logistic classifiers. The proposed model is found to perform better than the benchmark models or equally weighted ensembles.

Original languageEnglish
JournalJournal of Forecasting
DOIs
Publication statusAccepted/In press - 1 Jan 2019

Fingerprint

Memory Term
Stock Market
Ensemble
Neural Networks
Neural networks
Data storage equipment
Benchmark
Technical Analysis
Lasso
Model
Nonstationarity
Operating Characteristics
Ridge
Logistics
Weighting
Classifiers
Receiver
Classifier
Financial markets
Stock market

Keywords

  • deep learning
  • ensemble models
  • high-frequency trading
  • LSTM neural networks

Cite this

@article{f41b921e1b604cce9f47020ca1e28f0d,
title = "An ensemble of LSTM neural networks for high-frequency stock market classification",
abstract = "We propose an ensemble of long–short-term memory (LSTM) neural networks for intraday stock predictions, using a large variety of technical analysis indicators as network inputs. The proposed ensemble operates in an online way, weighting the individual models proportionally to their recent performance, which allows us to deal with possible nonstationarities in an innovative way. The performance of the models is measured by area under the curve of the receiver operating characteristic. We evaluate the predictive power of our model on several US large-cap stocks and benchmark it against lasso and ridge logistic classifiers. The proposed model is found to perform better than the benchmark models or equally weighted ensembles.",
keywords = "deep learning, ensemble models, high-frequency trading, LSTM neural networks",
author = "Svetlana Borovkova and Ioannis Tsiamas",
year = "2019",
month = "1",
day = "1",
doi = "10.1002/for.2585",
language = "English",
journal = "Journal of Forecasting",
issn = "0277-6693",
publisher = "John Wiley and Sons Ltd",

}

An ensemble of LSTM neural networks for high-frequency stock market classification. / Borovkova, Svetlana; Tsiamas, Ioannis.

In: Journal of Forecasting, 01.01.2019.

Research output: Contribution to JournalArticleAcademicpeer-review

TY - JOUR

T1 - An ensemble of LSTM neural networks for high-frequency stock market classification

AU - Borovkova, Svetlana

AU - Tsiamas, Ioannis

PY - 2019/1/1

Y1 - 2019/1/1

N2 - We propose an ensemble of long–short-term memory (LSTM) neural networks for intraday stock predictions, using a large variety of technical analysis indicators as network inputs. The proposed ensemble operates in an online way, weighting the individual models proportionally to their recent performance, which allows us to deal with possible nonstationarities in an innovative way. The performance of the models is measured by area under the curve of the receiver operating characteristic. We evaluate the predictive power of our model on several US large-cap stocks and benchmark it against lasso and ridge logistic classifiers. The proposed model is found to perform better than the benchmark models or equally weighted ensembles.

AB - We propose an ensemble of long–short-term memory (LSTM) neural networks for intraday stock predictions, using a large variety of technical analysis indicators as network inputs. The proposed ensemble operates in an online way, weighting the individual models proportionally to their recent performance, which allows us to deal with possible nonstationarities in an innovative way. The performance of the models is measured by area under the curve of the receiver operating characteristic. We evaluate the predictive power of our model on several US large-cap stocks and benchmark it against lasso and ridge logistic classifiers. The proposed model is found to perform better than the benchmark models or equally weighted ensembles.

KW - deep learning

KW - ensemble models

KW - high-frequency trading

KW - LSTM neural networks

UR - http://www.scopus.com/inward/record.url?scp=85065740152&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85065740152&partnerID=8YFLogxK

U2 - 10.1002/for.2585

DO - 10.1002/for.2585

M3 - Article

JO - Journal of Forecasting

JF - Journal of Forecasting

SN - 0277-6693

ER -