Design and Construction of Recruitment Screening Model in Personnel Management System Based on Decision Tree

Under the background of big data era, the traditional recruitment method can no longer adapt to the more and more severe employment situation, and there is an urgent need to improve the ability of the recruitment system to deal with massive data. This paper combines user image technology and decision tree algorithm to successfully optimize the design of recruitment screening models in personnel management systems. Specifically, this paper proposes a data mining-based user image talent label generation method, and uses the CART decision tree algorithm to construct an intelligent recruitment model, which achieves accurate recruitment screening. Taking the recruitment of editing positions as an example, this paper explores the construction of the talent portrait, and its results show that: different editing positions have obvious differentiation in the demand for professional knowledge of talents, and employers pay special attention to vocational skills and digital literacy in recruitment, and generally pay attention to soft skills. The constructed CART model’s training prediction value is more accurate than the linear model’s, with a better fitting effect, which is suitable for talent recruitment screening tasks. At the same time, compared with the algorithmic models such as logistic regression, plain Bayes, support vector machine, CART decision tree, and K-nearest neighbor classifier, the F1 value of the CART model reaches 0.857, and in the two models of whether the talent is delivered or not and whether the job recruiter is recognized or not, its AUC value is 0.704 and 0.787, respectively, which is better than the other models, which indicates that the recruitment screening model built in this paper can achieve better application results, and can assist HR to complete the recruitment process and improve work efficiency.