Available via license: CC BY-NC-ND 3.0
Content may be subject to copyright.
Using Neural Networks to Predict Wages Based on Worker Skills
Abstract
This paper describes the results of our Neural Network (NN) models that predict annual wages based on the combination and levels of 35 different skills possessed by wage earners. These models can estimate the value of skills for skill-based compensation systems. They can be used by employers to determine how much to compensate different combinations of skills and by employees to estimate what they should be paid. Finally, governments can use the models to support the development and analysis of labor policies. We collect and integrate official U.S. Government data on 35 general job skills with the annual wage data for over 900 standard occupations. The skills data is then used as inputs to train an artificial intelligence (Al) neural network (NN) models. The resulting NN models train to above 70 percent accuracy in predicting annual wage levels based on the combination and levels of 35 different skills. This research makes use of authoritative U.S. Government data in new ways that can be used to better understand the connections between general skills and their relationships to wages. The need for these types of analytical tools is all the more important as changes in the job market have been severely impacted by the COVID 19 pandemic and are increasingly at risk of being replaced by automation.
... Predicting salaries using Neural Networks based on skills is one of the most important attributes. [13]. Secondly, the Random Forest Algorithm can also be used for better accuracy. ...
The study uses regression models to examine AI model engagement in employee salary prediction. Salary prediction is significantly critical for both employees and employers. Employees try to get maximum benefit for the services they provide, and employers emphasize achieving organizational goals through optimized employee salaries. Inconsistency in employee salaries may cause organizations financial losses and organizational objectivity failure. Historically, salary was determined by historical data, market surveys, and personal judgment, which often resulted in inconsistencies and biases. With the availability of large datasets from Human Resource Management Systems and advanced machine learning algorithms, there is an opportunity to enhance the fairness of Salary Prediction. To overcome this problem in organizations, we have proposed a regression model for salary prediction with a promising accuracy rate of 99% with regression models. The methodology includes data processing steps including EDA, data standardization, feature correlation, and feature engineering to enhance the accuracy of the models. This study used Random Forest Regressor, Gradient Boosting Regressor, and Light Gradient Boosting Machine Regressor models for employee salary prediction. This research paper provides a valuable understanding of HR analytics for HR professionals and organizations for salary prediction of employees in an organization. Also, this research investigates the use of Machine learning algorithms to predict employee salaries while comparing employee performance and eliminating biases. The aim is to develop robust, data-enriched frameworks in HRMS for organizations for accurate and transparent salary prediction.
... Professional cultural appreciation refers mainly to the level of cultural knowledge of practitioners, which is the basic quality for practitioners to adapt and develop in their occupations [56]. Salary is generally divided into skill-based compensation [57], positionbased compensation, and competency-based compensation. Professional salary is one of the main factors influencing occupation differentiation and the social division of labor-occupation perception includes recognition and satisfaction. ...
Owing to the increasingly complex economic environment and difficult employment situation, a large number of new occupations have emerged in China, leading to job diversification. Currently, the overall development status of new occupations in China and the structural characteristics of new occupation practitioners in different cities are still unclear. This study first constructed a development index system for new occupation practitioners from five dimensions (group size, cultural appreciation, salary level, occupation perception, and environmental perception). Relevant data to compare and analyze the development status of new occupation practitioners were derived from the big data mining of China’s mainstream recruitment platforms and the questionnaire survey of new professional practitioners which from four first-tier cities and 15 new first-tier cities in China. The results show that the development level of new occupation practitioners in the four first-tier cities is the highest, and the two new first-tier cities, Chengdu and Hangzhou, have outstanding performance. The cities with the best development level of new occupation practitioners in Eastern, Central, and Western China are Shanghai, Wuhan, and Chengdu, respectively. Most new occupation practitioners in China are confident about the future of their careers. However, more than half of the 19 cities are uncoordinated in the five dimensions of the development of new occupation practitioners, especially those cities with middle development levels. A good policy environment and social environment have not yet been formulated to ensure the sustainable development of new occupation practitioners. Finally, we proposed the following countermeasures and suggestions: (1) Establish a classified database of new occupation talents. (2) Implement a talent industry agglomeration strategy. (3) Pay attention to the coordinated development of new occupation practitioners in cities.
... Among the prediction model, algorithms including multiple linear regression, random forest, neural network, decision tree, etc., are frequently used. Many scholars have conducted a lot of research on salary prediction, such as using neural networks to predict wages based on workers' skills [1], predicting the per capita wages of urban mining units based on grey theory [2], predicting job salaries based on random forest algorithm [3], and conducting employment salary forecast via KNN algorithm [4]. However, these studies considered heavily based on some specific salary influencing factors and recruitment requirements or certain industries, which lacked generality. ...
This paper aims to build a salary prediction model based on the resumes of candidates in a recruitment environment. Point-biserial correlation analysis and random forest feature importance ranking methods are employed for the paper to conduct feature selection after the dataset is cleaned and preprocessed. Then, OLS linear regression is adopted to analyze the features selected, and three different models, including random forest regression, decision tree, and ridge regression, are applied in prediction experiments, helping obtain results to be compared and analyzed based on RMSE and MAE. Finally, a stacking ensemble method can be used to integrate and fuse different models to build the final salary prediction model. This model definitely has practical reference significance for both candidates and recruiters.
Based on the existing forms of interaction between expert systems in evaluating labor market competencies, this research aims to conceptually describe the functioning of a neural network system for assessing new competencies (using a multilayer network with Adaline neurons) in the labor market through a graphical model. The system’s functioning is shown as a process using the BPMN 2.0 process modeling language. The proposed scheme highlights the interaction between labor market actors (employers) and educational organizations in Russia. The research also proposes a fundamental scheme for integrating expert councils of educational organizations into data mining processes, labor market competency assessment, and enhancing the accuracy of the proposed neural network system. The functioning of the neural network system is described within three modules: data mining, communication-driven, and document-driven modules. The model identifies three top-level processes and nine subprocesses. Each subprocess is provided with documentary and informational support. The role of decision-making components (university expert councils) is described as a link in the neural human–machine assessment system. The process of developing relevant educational programs based on the evaluation of data collected through data mining is outlined. This research formulates a fundamental scheme for the interaction between employers (labor market actors), university expert councils, and federal authorities within a unified information space. The authors propose a concept for developing educational programs using neural network IT. A business process for forming educational programs is developed as a graphical model, displaying actors, support, top and lower-level processes, and connections.
Research background: This article discusses how artificial intelligence (AI) is affecting workers' personal and professional lives, because of many technological disruptions driven by the recent pandemic that are redefining global labor markets. Purpose of the article: The objective of this paper is to develop a systematic review of the relevant literature to identify the effects of technological change, especially the adoption of AI in organizations, on employees’ skills (professional dimension) and well-being (personal dimension). Methods: To implement the research scope, the authors relied on Khan's five-step methodology, which included a PRISMA flowchart with embedded keywords for selecting the appropriate quantitative data for the study. Firstly, 639 scientific papers published between March 2020 to March 2023 (the end of the COVID-19 pandemic according to the WHO) from Scopus and Web of Science (WoS) databases were selected. After applying the relevant procedures and techniques, 103 articles were retained, which focused on the professional dimension, while 35 papers were focused on the personal component. Findings & value added: Evidence has been presented highlighting the difficulties associated with the ongoing requirement for upskilling or reskilling as an adaptive reaction to technological changes. The efforts to counterbalance the skill mismatch impacted employees' well-being in the challenging pandemic times. Although the emphasis on digital skills is widely accepted, our investigation shows that the topic is still not properly developed. The paper's most significant contributions are found in a thorough analysis of how AI affects workers' skills and well-being, highlighting the most representative aspects researched by academic literature due to the recent paradigm changes generated by the COVID-19 pandemic and continuous technological disruptions.
The relationship between personality and salary was investigated among 4,150 managers. Individuals at five different managerial levels completed a measure of the Big Five personality dimensions as part of a work-related psychological assessment. The validity of personality for predicting salary was examined separately by managerial level, sex, as well as by purpose of assessment (selection versus development). Results indicated that personality predicts managerial salaries with useful levels of validity and thus is valuable for predicting extrinsic career success. While there was no evidence for differential validity by sex or purpose of assessment, results differed across managerial levels, with stronger relationships among the lowest and highest managerial groups (i.e., supervisors and top executives) largely due to increased predictor and criterion score variability.
Purpose
The purpose of this study was to investigate multiple indirect Big Five personality influences on professionals’ annual salary while considering relevant mediators. These are the motivational variables of occupational self-efficacy and career-advancement goals, and the work status variable of contractual work hours. The motivational and work status variables were conceptualized as serial mediators (Big Five → occupational self-efficacy/career-advancement goals → contractual work hours → annual salary).
Design/Methodology/Approach
We realized a 4 year longitudinal survey study with 432 participants and three points of measurement. We assessed personality prior to the mediators and the mediators prior to annual salary.
Findings
Results showed that except for openness the other Big Five personality traits exerted indirect influences on annual salary. Career-advancement goals mediated influences of conscientiousness (+), extraversion (+), and agreeableness (−). Occupational self-efficacy mediated influences of neuroticism (–) and conscientiousness (+). Because the influence of occupational self-efficacy on annual salary was fully mediated by contractual work hours, indirect personality influences via occupational self-efficacy always included contractual work hours in a serial mediation.
Implications
These findings underline the importance of distal personality traits for career success. They give further insights into direct and indirect relationships between personality, goal content, self-efficacy beliefs, and an individual’s career progress.
Originality/Value
Previous research predominantly investigated direct Big Five influences on salary, and it analyzed cross-sectional data. This study is one of the first to investigate multiple indirect Big Five influences on salary in a longitudinal design. The findings support process-oriented theories of personality influences on career outcomes.
The rapid development and systemic integration of new technologies has reignited concerns about the implications of automation for the volume and quality of jobs. However, there is surprisingly little empirical work that explores the drivers, constraints and employment impacts of technology from a business perspective across different sectors. This research note briefly interrogates the relevant literature, with particular regard to the New Zealand context, and argues the case for multi-sector, longitudinal research.
We examine how susceptible jobs are to computerisation. To assess this, we begin by implementing a novel methodology to estimate the probability of computerisation for 702 detailed occupations, using a Gaussian process classifier. Based on these estimates, we examine expected impacts of future computerisation on US labour market outcomes, with the primary objective of analysing the number of jobs at risk and the relationship between an occupations probability of computerisation, wages and educational attainment.
"The career is dead - long live the career!" 1 Such is the mixed message regarding careers that we are carrying into the next millennium. The business environment is highly turbulent and complex, resulting in terribly ambiguous and contradictory career signals. Individuals, perhaps in self-defense, are becoming correspondingly ambivalent about their desires and plans for career development. The traditional psychological contract in which an employee entered a firm, worked hard, performed well, was loyal and committed, and thus received ever-greater rewards and job security, has been replaced by a new contract based on continuous learning and identity change, guided by the search for what Herb Shepard called "the path with a heart." In short, the organizational career is dead, while the protean career is alive and flourishing. In this special issue of The Executive we will examine the ways the career environment and the executive of the 21st century will shape the direction of careers in the years to come. In this opening paper, we will provide a brief overview of the emerging career landscape, for both organizations and individuals. Then we will turn to an overview of the papers in this Special Issue and then to the papers themselves.
How Much Are Your Skills Worth? Available online at
- Dice
- Com