RESUMEN
Consumer behavior variants are evolving by utilizing advanced packing models. These models can make consumer behavior detection considerably problematic. New techniques that are superior to customary models to be utilized to efficiently observe consumer behaviors. Machine learning models are no longer efficient in identifying complex consumer behavior variants. Deep learning models can be a capable solution for detecting all consumer behavior variants. In this paper, we are proposing a new deep learning model to classify consumer behavior variants using an ensemble architecture. The new model incorporates two pretrained learning algorithms in an optimized fashion. This model has four main phases, namely, data gathering, deep neural modeling, model training, and deep learning model evaluation. The ensemble model is tested on Facemg BIG-D15 and TwitD databases. The experiment results depict that the ensemble model can efficiently classify consumer behavior with high precision that outperforms recent models in the literature. The ensemble model achieved 98.78% accuracy on the Facemg database, which is higher than most machine learning consumer behavior detection models by more than 8%.