مطالعات منابع انسانی (Oct 2023)
Designing a Knowledge Graph with Data-Driven Algorithms to Optimize Matching People and Jobs and Ranking Skills
Abstract
Background & Purpose: The job market has recently undergone significant changes due to the factors such as digitalization, the tendency to work remotely, and rapid changes in market demands, leading to challenges such as reduced job tenure, an increase in the working population, and a widening gap between skills and jobs. Therefore, it is necessary to do studies to investigate these challenges using novel approaches and offer new solutions. Despite the considerable importance of this issue, few studies in the literature have focused on it, especially data-driven approaches that have not been fully studied and discussed. Therefore, in response to the aforementioned challenges, this research has been conducted to design a knowledge graph based on data-driven algorithms to provide an accurate model of the alignment between skills and jobs. The primary goal is to create a practical framework in accordance with the dynamic nature of the job market, which can be used to improve the connection between job seekers and available job opportunities, as well as to rank skills based on their importance in the job market, and to identify the most efficient job transitions.Methodology: In order to design the proposed knowledge graph, job and skill data were collected from international classifications using natural language processing methods, and job advertisement data were also gathered to ensure alignment with the current job market conditions. The Jaccard index was used for matching skills, preferential attachment, and Node2Vec algorithms for predicting knowledge graph edges, and Dijkstra's algorithm for finding the most efficient job transitions.Findings: This research presents a new method for designing a knowledge graph of skills and jobs, outlining three important applications: 1- Quantifying the relationship between skills and jobs, 2- Finding the most efficient job transitions using shortest path algorithms and calculating job similarities, and 3- A method for ranking the necessary skills of each job group based on the created knowledge graph.Conclusion: This knowledge graph is a powerful tool for analyzing and understanding the job market, utilizing algorithmic methods for data analysis. It enables a comprehensive match between job seekers and job opportunities and the relationship between skills and jobs will be predictable. Additionally, the knowledge graph can be used to investigate skill-based job similarities for career pathfinding and rank the unique skills of each job group for optimal learning and development of individuals.
Keywords
