Machine Learning in Forecasting Labor Market: A Systematic Literature Review

Rapid changes and advancements in science and technology continuously reshape the labor market. With the emergence of new intelligent systems, traditional work practices become obsolete, new occupations arise, while some existing roles decline. As a consequence, it becomes a crucial activity for the workforce to be adaptive to the labor market needs, to keep their expertise up-to-date and aligned with developing trends. However, analyzing labor market is challenging for an individual as it requires considering multiple, dynamic inputs in the process. Although recent studies have explored the use of Machine Learning (ML) for labor market forecasting, there remains limited clarity on the predictive models employed, the data sources utilized, and whether these approaches adequately address data inconsistency and biases, sequential stability, and explainability of predictions. This study presents a systematic review of the current state of research on the application of ML techniques in the labor market forecasting, published until November 2024, aiming to identify existing methodologies, and their applications. The review identifies commonly used predictive models, data sources, and forecasting outcomes. It further examines how researchers address data inconsistency, sequential stability of prediction, and model explainability. Our findings revealed that although a wide range of ML models and data sources are employed to forecast future skills, automation risks, and labor competition, most studies fail to adequately address critical concerns such as data inconsistency, model explainability, and sequential stability of prediction over time. These gaps highlight the limited reliability of current approaches and underscore the need for more transparent and temporally robust forecasting frameworks.