MatchingSem: Online recruitment system based on multiple semantic resources

Abstract

The growth of online recruitment has spurred the need for more effective automated systems. On the one hand, traditional approaches based on keyword-based matching techniques suffer from low precision, i.e. a large fraction of the systems' suggestions are irrelevant. On the other hand, the newer semantics-based approaches are penalized by limitations of the exploited semantic resources, namely semantic knowledge incompleteness and limited domain coverage. In this paper, we present an automatic semantics-based online recruitment system that reuses knowledge captured in multiple existing semantic resources to match between candidate resumes and job posts. In addition, we use statistical-based concept-relatedness measures to alleviate the problem of semantic knowledge incompleteness in the exploited resources. An experimental instantiation of the proposed system has been installed to validate its effectiveness in matching job applicants to job posts.

Semantics-based key concepts identification for documents indexing and retrieval on the web

“…For instance, in Wimalasuriya and Dou (2009), the authors demonstrated through experimental results that by using multiple ontologies the quality of the system's precision can be improved. Similarly, Kmail et al (2015) have confirmed the increase in the system's effectiveness when utilising multiple semantic resources for the purposes of matching resumes to their corresponding job posts. However, the authors in this work have also acknowledged the fact that even with the exploitation of multiple semantic resources, some entities were still unrecognised by any of the used resources.…”

“…With the growing interest in semantic resources, several recent approaches have been proposed for semantically analysing user queries and matching them at a semantics-based level to their related documents. However, these approaches either use a single semantic resource such as Lu et al (2015), , Han et al (2016), Selmi et al (2018), Boiński et al (2018) and Royo et al (2005) or multiple heterogeneous semantic resources such as Maree et al (2016), Vigneshwari and Aramudhan (2015), Shen and Lee (2018), Kmail et al (2015), Zhu and Iglesias (2018), Goldfarb and Le Franc (2017) and Wimalasuriya and Dou (2009). For instance, the system proposed in Royo et al (2005) maps query keywords to their corresponding synsets in WordNet ontology.…”

“…Recently, there has been attempts to exploit multiple semantic resources in specialised domains such as the recruitment domain (Kmail et al, 2015), biomedical information retrieval domain (Shen and Lee, 2018) and the information extraction domain (Wimalasuriya and Dou, 2009). For instance, in Wimalasuriya and Dou (2009), the authors demonstrated through experimental results that by using multiple ontologies the quality of the system's precision can be improved.…”

“…This was either because such entities have been introduced to the domain after the creation of the respective semantic resource/s or because domain experts did not include them due to time and efforts required to define each and every term in the respective domains of interest. In the work proposed in Kmail et al (2015), the authors have utilised a set of semantic resources in the e-recruitment domain to map keywords in job posts and resumes to their corresponding entities in the used ontologies. The authors have conducted a set of experiments to demonstrate that the exploitation of multiple semantic resources can lead to better assigning relevance scores between resumes and their corresponding job posts.…”

Increasing the efficiency of recruitment based on deep neural networks

“…Третий подход -использование семантики [17,18] для решения задачи классификации резюме. Исследование [6] выявляет повторяющиеся особенности в семействах должностей при помощи метода LDA [13,14] и кластеризации (K-means) [21,23] и использует их для характеристики каждой области. В нем также рассматривается относительная важность различных наборов навыков в этих группах должностей.…”