Research on Training Performance Prediction Model Based on LightGBM

Abstract

This study aims to build an accurate performance prediction model to assist the training and development of skilled workers. This article uses the advanced LightGBM algorithm and combines it with rich technical worker training data in a certain industry to predict trainees' training results. In the data processing stage, this article conducted detailed data cleaning to eliminate outliers and ensure the accuracy and reliability of the data. At the same time, by introducing new variables, this paper further enhances the feature expression ability of the model, thereby improving the accuracy of prediction. To further optimize the model performance, this article innovatively uses the particle swarm algorithm to fine-tune the parameters of LightGBM. Experimental results show that the optimized model performs well in predicting students' training performance, with not only high prediction accuracy but also good stability. This model is expected to become a powerful tool for the training and development of skilled workers, helping various industries to evaluate and improve training effects more efficiently, thereby better meeting the needs of modern production and promoting the improvement of the overall quality of the skilled worker team.