Polynomial Regression Framework for Predictive Salary Estimation of Non-Academic Personnel

Abstract

Salary prediction has become an essential application of data-driven analytics in human resource management, supporting organizations in establishing transparent, equitable, and evidence-based compensation strategies. Traditional salary determination methods often rely on manual assessment, subjective judgment, and limited statistical analysis, resulting in inconsistencies and reduced decision-making efficiency. This study proposes a Polynomial Regression Framework for Predictive Salary Estimation of Non-Academic Personnel, integrating machine learning principles with polynomial regression to model the nonlinear relationships among employee characteristics and salary outcomes. The framework emphasizes data preprocessing, feature engineering, polynomial transformation, model training, validation, prediction, and performance evaluation. Unlike linear regression, polynomial regression captures complex interactions among variables such as educational qualification, years of experience, employment duration, professional certifications, departmental responsibilities, and performance ratings. The proposed methodology synthesizes concepts from regression-based salary prediction, machine learning algorithms, and human resource analytics to develop a practical predictive model suitable for institutional environments. The study further discusses theoretical implications, implementation strategies, and limitations while highlighting how intelligent analytical frameworks improve workforce planning and compensation management. The proposed framework demonstrates that polynomial regression provides an interpretable, computationally efficient, and scalable solution for salary estimation where nonlinear relationships exist. The findings indicate that integrating machine learning techniques into human resource management enhances prediction reliability and organizational decision support while maintaining model transparency and operational feasibility (Bansal et al., 2021; Das et al., 2020; Yang, 2023). The importance of intelligent analytical frameworks for organizational decision optimization is also consistent with recent integrated predictive system architectures proposed by Geo Philip and Paulson (2025).