| Abstrakt: |
This study offers a comprehensive exploration of unemployment, spanning short-term, medium-term and long-term periods, highlighting the necessity for in-depth analysis and timely intervention. The primary goal is to develop machine learning models proficient in identifying the characteristics of individuals undergoing unemployment. Utilizing data from the European Social Survey, the research employs data processing techniques and diverse machine learning algorithms, including Logistic Regression, Random Forest, Ada Boost, LightGBM, and Gradient Boosting. The comparative analysis of unemployment across different durations reveals significant variations in predicting factors. Short-term unemployment is notably influenced by age. For those unemployed for twelve months or more, factors such as perceived equal opportunities and attachment to Europe become influential. In the context of unemployment lasting five years, variables related to happiness and marital status prove crucial, impacting motivation and potentially contributing to extended unemployment. This study marks the initiation of a comprehensive exploration into understanding the traits of unemployed individuals, extending beyond the individual level to incorporate macroeconomic considerations. The application of machine learning algorithms provides valuable insights into addressing this the societal challenge of unemployment. [ABSTRACT FROM AUTHOR] |
