PREDICTION OF TRANSITION PROBABILITIES FROM UNEMPLOYMENT TO EMPLOYMENT FOR TURKEY VIA MACHINE LEARNING AND ECONOMETRICS: A COMPARATIVE STUDY

Authors : Yasin KÜTÜK, Bülent GÜLOĞLU

Pages : 58-75

View : 19 | Download : 12

Publication Date : 2019-03-15

Article Type : Research Paper

Abstract :In this study, it is mainly aimed to predict transition probabilities of individuals who are previously unemployed and get employment or stay unemployed. In order to do that, Household Labor Force Surveys conducted in Turkey are merged and matched from 2004 to 2016. Information about individuals only consists of individual characteristics and qualifications since there should not be any informative clue about the present situation. To predict those probabilities, logistic regression analysis as econometric approach, a shallow neural network and machine learning classification algorithms are run in order to compare them. The results indicate that classification in machine learning is slightly better than logistic regression and shallow neural network. While XGBoost classifier and Random Forest get 67% accuracy, logistic regression can predict only 63% of an individual’s transition and shallow neural network forecasts 51%.
Keywords : Employment, Transition Probability, Machine Learning, Classification