ABSTRACT

The aim is to optimise the procedures and reduce the workload of human resource management (HRM), thereby increasing the working efficiency and improving system performance. Deep learning (DL) algorithms are employed to build a CC neural network (BPNN)-based HRM system model. Then, this model is optimised and simulated, whose performance is verified through comparisons with other models. The comparative simulation demonstrates that the proposed system model converges the fastest. Results of the Leave-One-Out (LOO) method also prove the fastest convergence and the best optimisation effect of the proposed system model over classic models. In particular, it can converge at about 60 epochs and provides an accuracy of about 88.72%, 2.76% higher than other models at tops. Regarding the prediction performance, the proposed system model presents an excellent fitting effect. Through experiments, the constructed model converges faster and makes predictions more accurately, providing an experimental reference for the operation and intelligent development of HRM systems in the economic field in the future.

摘要

目的是优化程序，减少人力资源管理（HRM）的工作量，从而提高工作效率，提高系统性能。采用深度学习 (DL) 算法来构建基于 CC 神经网络 (BPNN) 的 HRM 系统模型。然后对该模型进行优化和仿真，通过与其他模型的比较验证其性能。对比仿真表明，所提出的系统模型收敛速度最快。留一法（LOO）方法的结果也证明了所提出的系统模型相对于经典模型的最快收敛和最佳优化效果。特别是，它可以在大约 60 个 epoch 时收敛，并提供大约 88.72% 的准确率，比其他模型高 2.76%。关于预测性能，所提出的系统模型呈现出极好的拟合效果。通过实验，构建的模型收敛速度更快，预测更准确，为未来经济领域人力资源管理系统的运行和智能开发提供了实验参考。