An innovative approach to corporate HR training based on deep learning

Abstract This article constructs an enterprise human resource training system based on the DACUM method, aiming to solve the critical problems in enterprise training. The article improves the training effect by optimizing the training process, establishing an effective assessment mechanism, and perfecting the feedback and incentive mechanism. Aiming at the limitations of the DACUM method in training guidance, the article introduces a rough support vector machine to categorize the dimensions of competency factors. It constructs an employee effectiveness evaluation model by calculating the weights. Combined with data mining techniques, it provides decision support for managers. The results show that the optimized training system significantly improves employees’ innovative behavior, psychological capital and job performance, in which the regression coefficients of creative behavior and psychological capital are 0.593 and 0.404, respectively. The degree of explanation of job performance in the three dimensions of task, relationship, and dedication is 18.8%, 32.7%, and 30.9%, respectively, which verifies the effectiveness and practical value of the training system.