Extraction and Application of Indicators in the Urban Resource RecommendationProcess Using Structured and Non-Structured Data
The main objective of this research is to develop an approach capable of analyzing the information expressed in location-based social networks (LBSNs) in the context of recommending urban resources with the use of KDT in order to relate aspects related to information polarity and reliability of the profiles that broadcast them. Therefore, Web Mining approaches were used in the knowledge discovery and data analysis process. The extracted features were then applied to traditional recommendation algorithms, based on neighborhood and matrix factorization, in order to determine accuracy metrics with their use.