Skip to main navigationSkip to main content
The University of Southampton
Centre for Risk Research

The Great Divide in Financial Market Forecasting: Machine Learners vs. Financial Economists.  Event

Financial Markets
Time:
13:00 - 14:00
Date:
18 February 2016
Venue:
Room 4005 Building 4 (Law) Highfield Campus University of Southampton University Road Southampton SO17 1BJ

For more information regarding this event, please telephone Dr Ian Dawson on 023 8059 8094 or email I.G.Dawson@Soton.ac.uk .

Event details

Abstract: Financial time series forecasting is a popular application of machine learning methods. Previous studies report that advanced forecasting methods predict price changes in financial markets with high accuracy and that profit can be made trading on these predictions. However, financial economists point to the informational efficiency of these markets, which questions price predictability and opportunities for profitable trading. The objective of the paper is to resolve this contradiction. We achieve this by clarifying which factors explain the apparent success of machine learning methods in previous research, in order to examine the generalizability of such findings. To pursue these goals, we examine the efficiency of financial markets through the lens of methodological factors that characterize forecasting experiments. The results of a comprehensive forecasting simulation suggest that the predictability of a financial market and the feasibility of profitable model-based trading are significantly influenced by the maturity of the financial market, the forecasting method employed, the horizon for which it generates predictions and the methodology used to assess the model and simulate model-based trading. We find no evidence that indicators from the field of technical analysis increase the performance of data-driven forecasting models. Overall, our findings confirm that advanced forecasting methods can be used to predict with some accuracy, price changes in some financial markets. However, our results suggest that previous machine learning experiments may have led us to overestimate the degree of inefficiency prevalent in markets. In the light of these findings, we discuss whether results from machine learning based financial time series forecasting studies should lead us to question the prevailing view in the financial economics literature that financial markets are efficient.

Speaker information

Ming-Wei Hsu,Centre for Risk Research, University of Southampton. , Ming-Wei Hsu was awarded a BSc and a MSc in Computer Science from National Taiwan University. He then joined a software company, CyberLink, and worked as an engineer for 5 years. Subsequently, he was awarded a MSc in Accounting and Management from the Southampton Business School, University of Southampton. He is currently studying for a PhD at the Centre for Risk Research, University of Southampton. His research examines the effects of learning among traders in financial markets and uses spread trading data. His results suggest that traders increase both returns and volatility of returns as they gain experience, which leads to the decrease of the risk-adjusted investment performance measured by Sharp ratio. He is also exploring the extent to which it is possible to predict the direction of financial market indices using machine learning techniques, such as Support Vector Machine and Artificial Neural Networks.

Privacy Settings