Effective statistical methods in big data

This thesis aims to use big data in order to determine the features affecting the price of a product with the hedonic regression model and to estimate the contribution of each feature to the price by using robust regression estimation methods. For the analysis, the price and features of the laptop product groups were obtained from the big data source by using the web scraping method. Four different alternatives of the hedonic regression model are used to determine the features affecting the price of the laptops. The contribution of each feature to the laptop price is estimated by using the robust estimation methods (M and MM) and the Ordinary Least Squares (OLS) estimation method, and the resulting estimates are compared for both methods. By using robust methods that minimize the effect of outliers that can be observed in the data, it is aimed that the outliers in the data will have the least effect on the analysis result. In the framework of the data set used in the study, it is observed that the effective model is the Logarithmic Robust Hedonic Regression Model.