REINFORCEMENT LEARNING IN PREDICTIVE ANALYTICS FOR HUMAN BEHAVIOUR

ABSTRACT The development of the internet hasled to the digitization of data, opening up opportunities in big data. This digital footprint reveals extensive insights into what people read, how they behave and their involvement in various activities, shedding light on their likes and dislikes. The goal is to use predictive analytics to optimize the predicition of human behaviour for strategic insights and decision- making. Predictive analytics contains a range of statistical and analytical techniques designed to enable organizations to forecast future outcomes and events with a degree of certainty based on historical data and interactions. This process can be enhanced by employing self-learning algorithms and models that predict human behaviour from data-driven insights and refine their predictions based on feedback and results,which is the central process of Reinforcement Learning(RL). RL is poised to revolutionize the field of AI and aims to automate systems with an advanced understanding of the virtual world and tackle problems previously consideredintractable This paper starts with an introduction to the general field of RL, followed by an exploration of value- based and policy- based methods.It then delves into the core concept of predictive analytics and its essential components. We discuss the main algorithms of RL and their connection to human operant learning, demonstrating why RL-based predictive analytics is better than the traditional way of it. Proceeded by applications and benefits of using RL in predicting human behaviour . In conclusion, we propose the use of RL models in predictive analytics to predict human behaviour through in formatics and analytics approach, aiming to gain deeper insights into human behaviour to enhance decision-making and strategic insights. Keywords: Reinforcement learning, predictive analytics,big data, human prediction , self learning algorithms