Dynamic and Efficient Human Resource Management System Using an Improved Random Forest Model

Regarding the construction of a human resource management system and analysis of influencing factors, the traditional random forest (RF) algorithm has high computational complexity and large resource requirements when processing large-scale data, resulting in low system analysis efficiency and a lack of real-time analysis and dynamic adjustment. This paper constructs an improved RF model, collects multi-dimensional data from the internal database of the enterprise, and uses standardization and denoising techniques for preprocessing to ensure data quality and consistency; based on the traditional RF algorithm, a new feature weighting strategy is adopted, and the splitting criterion of the decision tree is optimized during construction. This paper introduces a dynamic adjustment mechanism to monitor data flow changes in real time and adjust the model parameters and training strategies promptly, so that the system can flexibly respond to different human resource management needs; it adopts a distributed computing framework to improve the training speed on large-scale data sets and optimize the system’s response time. The experimental results show that basic salary and bonus have the greatest impact on employee management, with weights of 0.18 and 0.14, respectively. The training time of the improved RF model tends to be 6[Formula: see text]s, and the memory usage tends to be 110[Formula: see text]MB after 50 training rounds, which has higher training efficiency and lower resource consumption. Under 5000 data streams, the response time of the improved RF model is 23[Formula: see text]s, which reflects a certain degree of real-time and dynamic adjustment capabilities; the experimental results prove the effectiveness of this paper in the construction of a human resource management system and the analysis of influencing factors.