Deep Fuzzy Ontology for Travel User Interest Discovery Based on Visual Shared Data in Social Networks

Abstract

In the present research, a novel system for travel interest discovery from visual shared data through social networks is discussed. The proposed Deep Fuzzy System is based on neural features extracted from well-known CNN architecture. GoogleNet to learn the inference system based on fuzzy ontology. Deep fuzzy ontology is a new framework that includes essentially two phases. The first consists of image conceptualization by existing objects found in shared images. In the second phase, we use the concepts issued from ImageNet classes to design our Tree-based ontology for travel interest discovery.To evaluate our deep fuzzy CNN ontology system, we construct a new database of visual shared data on Facebook coming from Sudanese citizens. Our proposed system has shown a very impressive result for travel Sudanese user’s interest.