Chatbot based on Deep Learning for Recommending Relevant Products

Luis Alberto Pachas-Santos, Hugo David Calderon-Vilca, Flor C. Cardenas-Mariño


Online shopping platforms are growing at an unprecedented rate around the world. These platforms are mostly based on search engines, which are still mostly knowledge base based and use matching keywords to find similar products. However, customers want a more interactive approach that is convenient and reliable for viewing related products. In this article, we propose a novel idea of ​​searching for products in an online shopping system using a chatbot model that uses Deep Learning which processes images and text to generate responses to the customer. A user can provide a picture or give a description of the product they are looking for, and they will be presented with similar products based on the images. The system of proposed recommendations is based on the recovery of images based on the content provided. The evaluation of the proposed model is divided into two parts: if it is done through images it generates 80.6% accuracy and if it is done through text it has 75% accuracy.


Relevant products, recommendation, chatbot, machine learning, neural networks, selling chatbot

Full Text: PDF (Spanish)