Finding the appropriate housing: A fuzzy-model-based recommender system

Abstract

When looking at housing concerns, the most frequent question posed by citizens finding a suitable site in agreement with their wishes. Nowadays, they are faced with cent housing offers through the internet. In this context, new decision support systems are claimed to help citizens choose the best one. Recommender systems are currently successful solutions to support users who adjust their preferences with information overload. However, to deal with human uncertainty, new challenges arise in recommender systems that have been little explored. This work proposes an approach to suggest the appropriate housing, based on the decent housing model proposed by the United Nations, in the framework of the Habitat III conference and citizen's wishes. An exploratory research approach was used. The process started picked up the data set from the platform PUC 1; after that, the data set was cleaning and predicting missing values using a fuzzy c-mean with neural networks; the model is trained with the membership function using c-mean. Finally, the recommendation starts from a person's tastes with the trained model, which assigns him a cluster with a degree of membership, proceeds with the filtering of seven neighborhoods, recorded in the memory of the recommended for future direct recommendations. The model tested could recommend housing that belongs to specific neighborhoods according to the users' preferences with some uncertainty.