Human-Augmented Prescriptive Analytics With Interactive Multi-Objective Reinforcement Learning

The rise of Artificial Intelligence (AI) enables enterprises to manage large amounts data in order derive predictions about future performance and gain meaningful insights. In this context, descriptive predictive analytics has gained a significant research attention; however, prescriptive just started emerge as the next step towards increasing maturity leading optimized decision making ahead time. Although machine learning for been identified one most important applications AI, up now, is mainly addressed with domain-specific optimization models. On other hand, existing literature lacks generalized models capable being dynamically adapted according human preferences. Reinforcement Learning, third paradigm alongside supervised unsupervised learning, potential deal dynamic, uncertain time-variant environments, huge states space sequential processes, well incomplete knowledge. paper, we propose human-augmented approach using Interactive Multi-Objective Learning (IMORL) cope complexity real-life environments need human-machine collaboration. process modelled way assure scalability applicability wide range problems applications. We deployed proposed stock market case study evaluate proactive trading decisions that will lead maximum return minimum risk user's experience available can yield combination.