Bootstrap Estimation in Quantile Regression Analysis

44877

One of the auxiliary analysis methods in regression analysis is the quantile regression method. There is no distributional assumption is required in the quantile regression model and it gives better estimates in the data sets of the structure where the outliers are estimated because it predicts the parameter coefficients depending on the various quantities. In addition, the quantile regression allows for the determination of the variance. In linear regression analysis, there are requirements such that data structure is suitable for the model.In this study; Linear regression and quantile regression methods are introduced, the differences between them are indicated. Information about the bootstrap method is given. In the application part, monthly Producer Price Index, 2 period delay and Expectation Questionnaire data between 2000-2017 were used. This data was used to increase the number of data at certain levels by Bootstrap method and to compare the results of Linear and Quantile Regression methods with Mean Absolute Deviation (MAD) and Root Means Square of Error (RMSE) to determine which method predicts the most suitable model. Results, Linear and Quantile Regression (50) methods show that the two methods with the closest and smallest values of MAD and RMSE predict the most suitable models.

16

Linear Regression , Quantile Regression , Bootstrap Estimation , RMSE

Erciyes University Journal Of The Institute Of Science an‎d Technology

25

Erciyes University Journal Of The Institute Of Science an‎d Technology