A New Ensemble Prediction Approach to Predict Burdur House Prices

Abstract

AbstractLooking at the hierarchy of needs, although housing is among the most basic physical needs of an individual, houses have also been used for a long time to generate income. However, determining the price of the house in the purchase and sale transactions is a challenging study since many independent variables must be examined. In order to overcome this difficulty, an ensemble estimation approach was developed in which many different prediction models were brought together in this study. In order to test the developed approach, a new data set was created by compiling the house sales advertisements in the province of Burdur. The data set samples created contain basic characteristics such as the square meter size of the houses, the number of rooms, and the age of the building. In addition to these qualities, the distances of the residence from educational institutions, hospitals, city center and other public institutions were also calculated and included in the data set in order to make the estimation process better. The data set obtained was used to train the developed approach, compare and test the predictions. The findings obtained as a result of the training and testing processes were shared.KeywordsHouse price predictionEnsemble algorithmRegression