Intelligent Deep Learning Systems for Predicting Customer Purchase Intentions in Phygital Retail Industry

The concept of phygital retail promotes the integration of digital and physical channels toward enhancing customer engagement, hence satisfaction. This research tried to know whether the intelligent deep learning system would be able to predict customer purchasing behavior in a phygital environment. In this paper, high data that is comprised of touchpoints-both digital and physical- were implemented into various deep learning models which included ANNs, CNNs, RNNs, and LSTM. In this research, it has been depicted that both RNN and ANNs become the best model due to the highest accuracy as well as maximum AUC values which range at 92.70% and 0.97, respectively. Another insight that has come out of this visualization study is how correlation matrices and distribution plots highlighted the need to marry digital copies of online metrics like time spent or visits with the ideal customer journey. The models, through advanced data preprocessing and feature engineering techniques, were highly predictive. This study fills the gap in existing literature by highlighting the integration of deep learning into phygital retailing offering actionable strategies for personalized marketing