Soft skills in career prediction: a machine learning approach for role forecasting

The increasing complexity of the job market, coupled with rapid technological advancements, has made career prediction and workforce planning more challenging than ever. Traditional methods of career counselling and workforce analytics are being enhanced through machine learning (ML) techniques that utilise datasets to predict career transitions, assess job suitability, and address skills gaps. This study explores the role of ML-driven predictive models in career forecasting, with a particular focus on integrating soft skills. The research findings highlight that Support Vector Machines (SVC) achieved the highest test accuracy (100%), though cross-validation adjusted this to 92% ± 3%, indicating potential overfitting. Neural Networks demonstrated high accuracy (96.77%) but incurred high computational costs, while Decision Tree performed well (93.55%) but showed susceptibility to overfitting. Additionally, the study revealed strong correlations between specific soft skills, such as problem-solving and collaboration, and job categories, underscoring the importance of these skills in predictive career modelling. The contribution of this study lies in its demonstration that ML models can effectively predict career pathways by incorporating soft skills assessment, thus enhancing traditional career recommendation systems and building a resilient workforce.