Career Satisfaction Prediction Using Machine Learning: Feature Importance and Performance Analysis

Career satisfaction is a dynamic and multifaceted phenomenon shaped by professional development, individual experiences, and career-related outcomes. This study aims to predict individuals’ career satisfaction using a machine learning-based regression framework with multidimensional features. The dataset consists of 400 observations, including demographic, academic, professional, and social attributes. Six machine learning algorithms, namely Linear Regression (LR), Decision Tree (DT), K-Nearest Neighbors (KNN), Support Vector Regression (SVR), Random Forest (RF), and Gradient Boosting (GB), were employed and comparatively evaluated. To enhance methodological rigor, potential data leakage risk was assessed through correlation analysis, and a restricted feature set was constructed by excluding variables highly correlated with the target variable. Model performance was evaluated using both cross-validation and train–test split approaches. Hyperparameter optimization was performed using GridSearchCV, and model performance was evaluated with coefficient of determination (R²), Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), Relative Absolute Error (RAE), and Root Relative Squared Error (RRSE) metrics. The results indicate that tree-based methods tend to outperform other models. After optimization, the GB model achieved the best performance on the held-out test set, with an R² value of 0.9213. Feature importance analysis using SHapley Additive exPlanations (SHAP) revealed that individual and career-related variables had a stronger influence on predictions. Overall, this study provides a methodologically transparent framework for predicting career satisfaction and supports data-driven decision-making in career-related contexts.