A Precision Advertising Strategy Based on Deep Reinforcement Learning

Abstract

Based on big data analysis, precision advertising fully meets the needs of users, and boasts a high application value. From the perspective of deep reinforcement learning (DRL), this paper attempt to develop a precision advertising strategy, capable of extracting effective features from massive advertising data and predicting advertising precision accurately and efficiently. Firstly, the advertising data were preprocessed, and organized into an advertising data sequence, in which the data are intercorrelated. In addition, the feature construction process was detailed. After that, a prediction model of advertising precision was developed in three steps, based on the Q-learning algorithm. The proposed strategy was found to be effective and accurate through experiments. The research results provide a reference for applying Q-learning to precision prediction in other fields.