AI Enabled Human Resource Analytics Framework for Workforce Performance Prediction using Functional Behaviour Data

Abstract

HR analytics is rapidly becoming essential to boost the productivity of organizations with data-informed HR management. This research suggests a framework for workforce performance prediction based on functional behaviour data that is assisted by artificial intelligence. The framework combines data normalization and NLP-based text cleansing to preprocess a diverse set of employee records, such as communication logs, task activities, and engagement data. SHAP-based feature importance selection is used to identify the most important behavioural attributes that impact workforce performance and enhance the model interpretability.

A Graph Neural Network with a Transformer-based classifier is designed and integrated with PyTorch and PyTorch Geometric to record complex interactions between employees and temporal patterns of their behaviors. The graph-based architecture captures the relationships among the nodes, while the Transformer mechanism captures the long-term dependencies in the activities of the workforce.

The experimental results show that the model is accurate and the features are easily interpretable, and the classification results are well and significantly superior to the conventional methods. The suggested framework facilitates intelligent human resource management choices, optimisation of human resources, development planning for employees and improving organizational performance in a changing work environment.