An Evaluation Framework for Machine Learning and Data Science-Based Financial Strategies: A Case Study-Driven Decision Model

Big data and computational technologies are increasingly important worldwide in asset and investment management. Many investment management firms are adopting these data science (DS) methods and technologies to improve performance across all investment processes. A good question is whether we can make better decisions in developing quantitative strategies. Therefore, the main objective of this research was to develop a multicriteria assessment framework and scoring decision support system to evaluate quantitative investment strategies that apply machine learning (ML) and DS techniques in their research and development. Subject matter experts will assess all framework perspectives from a systematic literature review to approve their reliability. The perspectives consist of <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">economic and financial foundations</i>, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">data perspective</i>, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">features perspective</i>, <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">modeling perspective</i>, and <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">performance perspective</i>. The research methodology applied was the hierarchical decision model (HDM) to provide a 360° view of the quantitative investment strategy and improve and generalize the concept to other asset classes and regions. This study accomplished a rigorous integration of an extensive literature review connecting DS, ML, and investment decision-making in developing quantitative investment strategies. As a result, the major contribution of this study is the comprehensive examination, which included identifying and quantifying perspectives and criteria. The results, while limited indicated significant gaps in strategies examined and therefore generated critical knowledge to improve ML/DS-driven investment strategies, which are valuable for financial companies and policymakers.