Prediction of Unemployment Rates in Turkey by k-Nearest Neighbor Regression Analysis

Abstract

Unemployment is a growing global problem. It provokes problems socially, psychologically and economically, considering it directly affects how people live in aspects of needs and hobbies, basically in the simplest meaning of living. Turkey is within the scope of economically growing countries and unemployment is one of the critical negative factors for targeted growth. In order to overcome this problem, unemployment rates can be estimated for the future with the available data. In this study, unemployment rates in Turkey are estimated by k-nearest neighbor regression (kNNR). For this purpose, a new dataset is created from the features that are likely to affect the unemployment rate. The performance of the kNNR algorithm is compared with two different machine learning algorithms which are ridge regression and linear regression. From the experimental results, it is observed that the kNNR method (coefficient of determination (R <sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">2</sup> ) = 0.7498) outperformed the compared methods. The results show that kNNR algorithm can be used effectively in this area.