ABSTRACT
Our goal is to improve the accuracy of labor analysis by utilizing Random Forests and Novel Convolutional Neural Networks (CNN) to provide instruments for identifying and stopping labor exploitation. Reliability was confirmed by using the G*Power test (α=0.05, power=0.85) to assess these techniques on two sets of 100 samples each. CNN outperformed Random Forest (96.33%) in accuracy, achieving 97.27%). A significant p-value of 0.001 was shown by an independent sample t-test, confirming the statistical difference between the approaches. The promise of advanced machine learning for labor analysis is shown by these findings. In order to promote ethical workplaces, a focus on model interpretability and ethical concerns is made. Subsequent investigations will concentrate on expanding the variety of datasets and tackling real-world implementation issues.
