Prediction model of vocational technical training effect based on machine learning

Data mining and machine learning can provide guidance for evaluation in the training field. Firstly, this paper applies the basic principles of Decision Tree (DT) and Random Forest (RF), and performs category weighting and output optimization on them to pay more attention to a few categories to improve the prediction accuracy of the model to analyze the evaluation data in the training field. Secondly, the correlation degree analysis is carried out on the discretized data, and the confidence matrix between the features is obtained. Finally, the key parameters such as DT pruning and the number of RF base classifiers are analyzed, and the performance of the two classifiers is compared by using 10-fold cross-validation. The binary classification accuracy of the two classifiers is 87.72% and 98.5% respectively under the optimal hyperparameters, which provides quantitative data support for evaluating students’ future development.