House rent estimation with machine learning algorithms: An application in Ankara

Today, with the effect of the pandemic, the need for housing is increasing day by day. During this period, people spend most of their days at home. The real estate market is one of the areas where machine learning can be applied to optimize and predict price with high accuracy. House price determination is a vital model in decision making for clients, where many parameters can be taken into account to predict the price of the desired house. Apartment rental prices are affected by several factors. In this thesis, a data set containing the rental price and different features of the apartments in Ankara Yenimahalle district will be examined, and it has been tried to analyze the different features of an apartment and to predict the rental price depending on multiple factors with the help of machine learning algorithms. Linear Regression, Lasso, Ridge, Random Forest, Xgboost algorithms were used. Among these algorithms, the best high accuracy algorithm was searched by changing the parameters. The algorithms used were compared with performance criteria such as MAE, MSE, RMSE, R2. Random Forest algorithm gave the best results in MAE, MSE, RMSE and R2 values.