Task recommender system using semantic clustering to identify the right personnel
| Published date | 13 May 2019 |
| DOI | https://doi.org/10.1108/VJIKMS-08-2018-0068 |
| Pages | 181-199 |
| Date | 13 May 2019 |
| Author | Prafulla Bafna,Shailaja Shirwaikar,Dhanya Pramod |
| Subject Matter | Information & knowledge management,Knowledge management,Knowledge management systems |
Task recommender system using
semantic clustering to identify the
right personnel
Prafulla Bafna
Department of Computer Studies, SICSR, Pune, Maharashtra, India, and
Shailaja Shirwaikar and Dhanya Pramod
Symbiosis International University, Pune, Maharashtra, India
Abstract
Purpose –Text mining is growing in importanceproportionate to the growth of unstructured data and its
applications are increasing day by day from knowledge management to social media analysis. Mapping
skillset of a candidateand requirements of job profile is crucial for conducting new recruitmentas well as for
performing internal task allocation in the organization. The automation in the process of selecting the
candidates is essential to avoidbias or subjectivity, which may occur while shuffling through thousandsof
resumes and other informative documents. The system takes skillset in the formof documents to build the
semantic space and then takes appraisals or resumes as input and suggests the persons appropriate to
complete a task or job position and employees needing additional training. The purpose of this study is to
extend the term-documentmatrix and achieve refined clusters to produce an improvedrecommendation. The
study also focuses on achievingconsistency in cluster quality in spite of increasing size of data set,to solve
scalabilityissues.
Design/methodology/approach –In this study, a synset-based documentmatrix construction method
is proposed where semanticallysimilar terms are grouped to reduce the dimension curse. An automated Task
Recommendation System is proposed comprising synset-based feature extraction, iterative semantic
clusteringand mapping based on semantic similarity.
Findings –The first step in knowledge extraction from the unstructuredtextual data is converting it into
structured form either as Term frequency–Inversedocument frequency (TF-IDF) matrixor synset-based TF-
IDF. Once in structured form, a rangeof mining algorithms from classification to clustering can be applied.
The algorithm gives a better feature vector representation and improved cluster quality. The synset-based
grouping and feature extraction for resume data optimizes the candidate selection process by reducing
entropyand error and by improving precision and scalability.
Research limitations/implications –The productivity of any organization gets enhanced by
assigning tasks to employees with a right set of skills. Efficient recruitment and task allocation can not
only improve productivity but also cater to satisfy employee aspiration and identifying training
requirements.
Practical implications –Industriescan use the approach to support different processes relatedto human
resourcemanagement such as promotions, recruitment and training and,thus, manage the talent pool.
Social implications –The task recommender system creates knowledge by following the steps of the
knowledge managementcycle and this methodology can be adopted in other similar knowledge management
applications.
Originality/value –The efficacy of the proposed approachand its enhancement is validated by carrying
out experiments on the benchmarkeddataset of resumes. The results are compared with existing techniques
and show refined clusters. That is Absoluteerror is reduced by 30 per cent, precision is increased by 20 per
cent and dimensions are loweredby 60 per cent than existing technique. Also, the proposed approachsolves
issue of scalabilityby producing improved recommendationfor 1,000 resumes with reduced entropy.
Keywords Clustering, Synset, Phrase, Silhouette coefficient, The feature vector
Paper type Research paper
Task
recommender
system
181
Received13 August 2018
Revised5 October 2018
8 December2018
Accepted15 January 2019
VINEJournal of Information and
KnowledgeManagement Systems
Vol.49 No. 2, 2019
pp. 181-199
© Emerald Publishing Limited
2059-5891
DOI 10.1108/VJIKMS-08-2018-0068
The current issue and full text archive of this journal is available on Emerald Insight at:
www.emeraldinsight.com/2059-5891.htm
1. Introduction
The success of any organization depends on the competence and superiority of its
employees. It becomes essential to choose the correct people and place them in the right
place. One needs to identify efficient methods of recruitment and task allocation. The
recruitment process is conducted by the HumanResource (HR) department of the company
to find out and attract capable applicants. The process provides the pool of potentially
skilled candidatesfor the job.
It starts with determining the present and future requirements of the organization in
conjunction with its personnel planning and job analysis activities. The HR team needs to
identify and prepare a pool of potential job applicants who will be appropriate candidates
(Derous and Ryan, 2016). The systematically conducted process increases the effectiveness
of individuals in the organizationsboth on the short- and long-term basis.
Task allocation is an internalactivity carried out in an organization in which workloads
associated with the project are distributed among the available employees. Task allocation
is based on the suggestion of senior employees who have experienced a successful task
completion from their team members. Sometimes employees also indicate their aspirations
or the tasks at which they are comfortable. The accomplishments and aspirations are
usually spelled out inself-appraisal forms and can be focused while allocating the task.
HR department uses resumes of aspiringcandidates and self-appraisal forms of existing
employees in their decision-makingprocess. Both being unstructured documents, it requires
efforts to extract the right information. Text mining techniques like clustering, feature
extraction and so on can be beneficial.In text mining, major terms (frequently occurring) are
considered as features representing a document (Bafna et al., 2016;Sun and Vasarhelyi,
2018). Clustering helps in grouping documents based on the similarity between the terms
present in the document. Semanticsimilarity can be emphasized further by making groups
of synonyms, meronyms and so on. This paper presents a holistic approach that combines
recruitment and task allocation that is assisted by mining initiatives for improving the
effectiveness of the decisionprocess.
A Synset-based Task Recommender System is proposed, for mapping skillset of an
employee to the job requirements. It is the mapping of expertise of the employees, to the
competencies required to perform the task effectively and efficiently. The employees who
are not able to fulfill the requirement of that task because of the absence of skillset, are
recommended training so that they will be eligible to perform the next similar project task.
The next section presents the background and related work. Task recommender system is
explained in the third section.The fourth section presents the experimental workcarried out
to prove the efficacy of the system. The paper ends with implications and limitations
followed by a conclusion.
2. Background and related work
The productivity of the employees heavily depends on the effective utilization of their
inherent competencies and expertise and focused efforts to impart training in skills and
technologies requiredfor the growth of the organization. HR department acquires, develops,
uses and maintains employees. Acquiringthe right men for the right job at the right time in
the right quantity, developing the right kind of training, using the selected workforce and
maintaining the workforce are the organizational objectives of the HR department. Thus,
choosing the right set of employees for the given task is the first step in this direction(Horne,
2016).
VJIKMS
49,2
182
To continue reading
Request your trial