Modelos de agrupamento e classificação para os bairros da cidade do Rio de Janeiro sob a ótica da Inteligência Computacional: Lógica Fuzzy, Máquinas de Vetores Suporte e Algoritmos Genéticos

Abstract

Since 2011, events of big repercussion have been taken place (and will continue to take for a period of time) in the city of Rio de Janeiro such as the conference Rio+20 of The United Nations and sporting events of world importance (World Cup, Olympic and Paralympic Games). These events allow the attraction of financial resources for the city, the creation of jobs, infrastructure improvements and the increase in real state value of buildings and territorials. By opting for a residential property in a particular neighborhood, a person does not only evaluates the property, but also the urban facilities available in the locality. In this context, it was possible to define a linguistic qualitative interpretation inherent in the neighborhoods of the city of Rio de Janeiro, integrating three Computational Intelligence techniques for the assessment of benefits: Fuzzy Logic, Support Vector Machine and Genetic Algorithms. The database was built with information from the web and government institutes highlights the cost of residential real estate, benefits and weaknesses of the city's neighborhoods. Initially, it was implemented Fuzzy Logic as an unsupervised clustering model, through the Ellipsoidal Rules by the Extension Principle using the Mahalanobis Distance, setting up as an inferential form the linguistic designation groups (Good, Regular and Bad) according with twelve urban characteristics. From this discrimination, it became tangible to use the integrated Support Vector Machine to Genetic Algorithms as a supervised method, in order to search/select the smallest subset of the variables in the group that best classify the neighborhoods (Principle of Parsimony). The analysis of error rates has enabled choosing the best classification model with a reduction of the space of variables, resulting in a subset that contains information about: HDI, number of bus lines, educational institutions, the average m² value, free outdoor spaces, entertainment places and crimes. The model that combined the three Computational Intelligence techniques put in hierarchy neighborhoods of Rio de Janeiro with acceptable error rates, collaborating on making of decisions for the purchase and sale of residential properties. When it comes to public transport in the city, it was revealed that the road network is still the priority