Deep Learning-Based Predictive Analytics for Intelligent Enterprise Management

Abstract

The problem of reference value in enterprise human resource management performance evaluation data is proposed, and the enterprise human resource management performance evaluation method based on machine learning is proposed. This paper builds an enterprise human resource management performance evaluation system from the three aspects of performance evaluation objectives, subjects and objects. HRM performance measures are chosen and measured in a standardized manner. The organization's employee-related data is analyzed using machine learning algorithms to create an intelligent predictive analytics framework for enterprise management. To enhance the quality and reliability of the data, the methodology proposed is data cleaning, removal of duplicates, noise filtering, One-Hot Encoding, Min-Max scaling, ADASYN data balancing and stratified data splitting. Random Forest (RF) and Extreme Gradient Boosting (XGBoost) methods are applied for predictive analysis and accuracy, precision, recall, and F1 Score are used to verify the performance of the methods. The experimental results show that the proposed RF shows 99.99% accuracy (ACC), 99.97% precision (PRE), 99.98% recall (REC), and 99.96% F1-score (F1), while the XGBoost shows 99.55% ACC, 99.96% PRE, 99.98% REC, and 99.97% F1. The findings show that the proposed models have shown excellent predictive performance and reliability, which can be used to help make decisions based on data and enhance the intelligent enterprise human resource management system.