Employment trend prediction and innovative talent training strategy optimization using naive Bayes classifier

Traditional methods for predicting employment trends primarily focus on graduates’ personal data, such as academic performance and professional background, often neglecting the influence of the macroeconomic environment on the job market. This oversight leads to significant deviations between predicted outcomes and actual employment trends, compromising prediction accuracy. This study proposes a novel approach using the naive Bayes classifier (NBC) to integrate multi-dimensional data, including historical employment information, graduates' personal data, and macroeconomic indicators, to enhance the precision of employment trend predictions. A dataset comprising employment data from 810 graduates across three higher vocational colleges in a specific region, combined with macroeconomic indicators, was constructed. Data preprocessing techniques, such as missing value filling and feature standardization, were employed to ensure data quality. Feature selection was performed using univariate linear regression to eliminate irrelevant variables and retain highly correlated features. The NBC-based model, augmented with semi-supervised classification to expand the training dataset, demonstrated superior performance, achieving an average error rate of 2.76% in predicting employment across five workplaces and a 95.68% accuracy in salary level prediction. The model efficiently processed 10,000 student records in 35.28 milliseconds. These results validate the effectiveness of NBC in employment trend prediction and provide a robust foundation for optimizing innovative talent training strategies.