Survey of Explainable Artificial Intelligence in Financial Forecasting

This document is a survey focused on Explainable Artificial Intelligence (XAI) in the context of financial time series forecasting. It categorizes various XAI approaches developed and published from 2018 to 2024, addressing the complexity of AI models that often hinders user trust, particularly in high-risk domains like finance. The survey defines key concepts, differentiates between explainability and interpretability, and discusses the significance of these distinctions in practical applications. It outlines the methodology used for categorization and includes a rigorous taxonomy of XAI approaches, along with their applications in the finance industry. The paper emphasizes the critical need for clear communication regarding AI decision-making processes to foster trust and regulatory compliance. Furthermore, it serves as a guide for data scientists and financial professionals seeking to implement XAI in predicting financial time series, providing an understanding of the merits and limitations of various XAI methodologies. This survey contributes to the responsible application of AI in finance while aligning with ethical standards.