Enhancing Enterprise Human Resource Management Through Intelligent Strategies Using Hybrid Deep Learning Models

Abstract

Introduction: Enterprise Human Resource Management (HRM), which tracks, evaluates, and improves employee performance, is essential to the expansion of a firm. However, conventional techniques like decision trees and linear regression frequently fall short in identifying intricate, non-linear relationships in employee data, which reduces their usefulness for making decisions in real time.
Objective: The research aims to develop an intelligent, accurate, and scalable model for forecasting employee performance using a hybrid Deep Learning (DL) approach called the Intelligent Water Drops Driven Dynamic Long Short-Term Network (IntWD-DynLSTN).
Methods: A real-world HR dataset that includes employee information like task completion rates, training hours, attendance, and performance ratings is used to train the algorithm. Preprocessing included encoding categorical data, using Z-score normalization, and addressing missing values by imputation. High-level characteristics were extracted from the structured HR data using Convolutional Neural Networks (CNN). These attributes were subsequently fed into a Dynamic Long Short-Term Network (DynLSTN) to identify sequential patterns in monthly employee performance. Finally, model hyperparameters were adjusted using the Intelligent Water Drops (IntWD) technique to enhance generality and accuracy.
Results: Experimental results show that the proposed IntWD-DynLSTN model achieves higher prediction accuracy (0.98), precision (0.97), recall (0.97), and F1-score (0.98) compared to traditional and baseline methods.
Conclusions: A scalable and dependable method for forecasting employee performance is provided by the suggested hybrid DL methodology. In dynamic organizational settings, it gives HR managers a strong tool for data-driven decision-making, facilitating quick interventions and efficient workforce management.