Interval Forecasting of Financial Time Series by Accelerated Particle Swarm-Optimized Multi-Output Machine Learning System

Jui-Sheng Chou,Thuy-Linh Le,Dinh-Nhat Truong

doi:10.1109/access.2020.2965598

Abstract

By providing a range of values rather than a point estimate, accurate interval forecasting is critical to the success of investment decisions in exchange rate markets. This work proposes a sliding-window metaheuristic optimization for interval-valued time series forecasting using multi-output least squares support vector regression (MLSSVR). The hyperparameters in MLSSVR are finetuned using an accelerated particle swarm optimization algorithm to yield the best predictions and fastest convergence. The proposed system has a graphical user interface that is developed in a computing environment and functions as a stand-alone application. The system is validated using stock prices as well as exchange rates and outputs are compared with published results. Finally, the proposed interval time series prediction method is tested in two case studies; one involves the daily Australian dollar and Japanese yen rates (AUD/JPY) and the other involves US dollar and Canadian dollar rates (USD/CAD). The proposed model is promising for interval time series forecasting.

Highlights

The foreign exchange market is a global financial market for trading currencies
INTERVAL TIME SERIES MODELING AND FORECASTING 1) SLIDING-WINDOW TIME SERIES ANALYSIS AND PHASE SPACE RECONSTRUCTION Owing to huge daily trading volumes, historical data from a foreign exchange market may be difficult to manage and its manipulation may result in a large computational burden
Yang (2010) [40] developed accelerated particle swarm optimization (APSO) as a simpler version of PSO to improve upon classical particle swarm optimization for finding a global solution and increasing the convergence rate

Summary

INTRODUCTION

The foreign exchange (forex) market is a global financial market for trading currencies. Accurate exchange rate forecasting provides a basis for making decisions about financial investments. Financial market forecasting based on a time series represents a means of providing information and knowledge to support a subsequent decision [6]. Several wellknown statistical models can be used in time series forecasting [7]–[10] Conventional modeling techniques, such as the Box-Jenkins autoregressive integrated moving average (ARIMA), are not adequate for financial market forecasting [10], [11]. This work develops a hybrid forecasting system using a user-friendly interface to predict exchange rates. The final section draws conclusions and makes recommendations concerning future researches

LITERATURE REVIEW

ACCELERATED PARTICLE SWARM OPTIMIZATION

PERFORMANCE EVALUATION METHODS

SYSTEM APPLICATIONS

Findings

CONCLUSIONS AND RECOMMENDATIONS