Sentiment Analysis and Image Classification in Social Networks with Zero-Shot Deep Learning: Applications in Tourism

Social media is one of the data sources that could provide more information or potential knowledge in almost any field of application. One of the main challenges of machine learning and big data is to solve the difficulty involved in the identification, classification, and, in general, the processing of all this data to extract useful information for a specific field. In this work, we propose a methodology for the detection of tourist places of interest through the combined use of images and text from social networks. For that purpose, we will be assisted by pre-trained neural networks for image classification and sentiment analysis. The result is frequency information of types of places according to a tourism-specific taxonomy combined with user sentiment indicators, which is potentially relevant information for tourism analysts.