Machine learning based forecasting of significant daily returns in foreign exchange markets
This work addresses the problem of improving forecasting accuracy for significant currency fluctuations, which is important for quantitative analysts and traders, but it is incremental as it applies existing outlier detection methods to a new application area.
The paper tackled forecasting significant daily returns in foreign exchange markets by applying machine learning algorithms, including a novel use of outlier detection methods, and found that outlier detection methods substantially outperform traditional techniques, with PKDE achieving the best results across various conditions.
Asset value forecasting has always attracted an enormous amount of interest among researchers in quantitative analysis. The advent of modern machine learning models has introduced new tools to tackle this classical problem. In this paper, we apply machine learning algorithms to hitherto unexplored question of forecasting instances of significant fluctuations in currency exchange rates. We perform analysis of nine modern machine learning algorithms using data on four major currency pairs over a 10 year period. A key contribution is the novel use of outlier detection methods for this purpose. Numerical experiments show that outlier detection methods substantially outperform traditional machine learning and finance techniques. In addition, we show that a recently proposed new outlier detection method PKDE produces best overall results. Our findings hold across different currency pairs, significance levels, and time horizons indicating the robustness of the proposed method.