Evaluating machine learning classification for financial trading: An empirical approach

作者：

Highlights：

• Low-complexity machine learning models are used trade in the FOREX market.

• A six year trading simulation in USDJPY, EURGPB and EURUSD are assessed.

• Periodic retraining, number of attributes and retraining set size are varied and studied.

• Middle range accuracies are obtained with high financial returns in the long term.

摘要

•Low-complexity machine learning models are used trade in the FOREX market.•A six year trading simulation in USDJPY, EURGPB and EURUSD are assessed.•Periodic retraining, number of attributes and retraining set size are varied and studied.•Middle range accuracies are obtained with high financial returns in the long term.

论文关键词：Trading,Financial forecasting,Computer intelligence,Data mining,Machine learning,FOREX markets

论文评审过程：Received 9 February 2015, Revised 8 January 2016, Accepted 12 January 2016, Available online 2 February 2016, Version of Record 19 February 2016.

论文官网地址：https://doi.org/10.1016/j.eswa.2016.01.018