A Holistic Auto-Configurable Ensemble Machine Learning Strategy for Financial Trading

摘要: Financial markets forecasting represents a challenging task for series of reasons, such as the irregularity, high fluctuation, noise involved data, and peculiar unpredictability financial domain. Moreover, literature does not offer proper methodology to systematically identify intrinsic hyper-parameters, input features, base algorithms strategy in order automatically adapt itself chosen market. To tackle these issues, this paper introduces fully automated optimized ensemble approach, where an feature selection process has been combined with automatic machine learning strategy, created by set classifiers hyper-parameters learned each marked under consideration. A experiments performed on different real-world futures demonstrate effectiveness approach regard both Buy Hold baseline several canonical state-of-the-art solutions.