Big Data, Hadoop and Spark for Employability: Proposal

Architecture

Aniss Qostal

, Aniss Moumen

and Younes Lakhrissi

Intelligent Systems, Georesources and Renewable Energies Laboratory,

Sidi Mohamed Ben Abdellah University, FST Fez, Fez, Morocco

Laboratory of Engineering Sciences, National School of Applied Sciences, Ibn Tofaïl University, Kenitra, Morocco

Keywords: Big Data, Data Analytics, Hadoop, Spark, Employability.

Abstract: The application of big data and data analytics has reached all aspects of life, from entertainment to scientific

research and commercial production. Mainly by taking advantage of the explosion of data at an unprecedented

rate attain the level of exabytes per day. On the other hand, it benefits from the sophisticated analytics

approaches that have been given new manners to translate the raw data into solutions and even into predictions

for complicated situations. This paper aims to discover the application of big data, data analytics and technical

architectures based on the Hadoop and Spark ecosystems to build employability solutions. Beginning with a

literature review of previous works and proposed solutions to draw a roadmap towards new approaches and

enhance the recruitment process for youth people.

1 INTRODUCTION

Today, employability is considered as a significant

concern for different entities involved in the

elaboration of the Labour Market Policies

(LMP)(Caliendo, 2016), with the primary goal to

enhance the employment opportunities and facilitate

the integration of job seekers and especially the youth

generation in the workplace; by improving the

matching process between job offers (vacancies) and

job seekers (i.e. the unemployed). Furthermore, the

basic philosophy behind the LMP is to supervise and

monitor the global atmosphere of the employment

environment and guaranty the necessary conditions

for good wellness in the workplace and ensure the

development of the workforce for the future

competencies required. Currently, according to the

reports of the international labour organization ILO in

2020, the global rate of unemployment has shown a

slight increase of 6.1% compared to 5.37% in 2019,

mainly due to the COVID-19 pandemic. The same

report has indicated the youth generation as the most

suffering layer with a rate of unemployment reaching

https://orcid.org/0000-0003-1016-8467

https://orcid.org/0000-0001-5330-0136

https://orcid.org/0000-0002-2537-8283

the highest level during a decade with 8.7% (Ryder

and Director-General, 2020).

On the other hand, we found many disciplines,

each in their domain of competence, try to find and

give their vision on the possible solutions for the

employability problems. Therefore, one of the

prominent and promising fields under investigation is

the use of the big data concept on employability and

the marketplace. Firstly as a way to treat data coming

from various sources related to employability.

Secondly, to support and enhance projects related to

the development of the labour market and facilitate

the integration of youth people.

In this paper, we will conduct an exploratory

study on the research carried out on the application of

big data for employment. Firstly, to discover

fundamental approaches developed on the

employability context, focusing on technical

architectures proposed based on Hadoop, Spark and

analytics with the aim to find solutions, propositions

and limitations and finally give answers to questions

like:

260

Qostal, A., Moumen, A. and Lakhrissi, Y.

Big Data, Hadoop and Spark for Employability: Proposal Architecture.

DOI: 10.5220/0010732200003101

In Proceedings of the 2nd International Conference on Big Data, Modelling and Machine Learning (BML 2021), pages 260-266

ISBN: 978-989-758-559-3

1) Can big data provide employment

Solutions?

2) Are there any approaches, initiatives or

projects launched to exploit big data in the

employability context?

3) Are there any initiatives in the developing

countries, especially the Moroccan context?

The results obtained from the previous researches

can be the base for proposal architecture and roadmap

for further studies.

2 MANUSCRIPT PREPARATION

2.1 Context

The application of big data in the employability

context benefited from the digitalization happened

last decades and from the growth of data generated

from different sources related directly or indirectly to

the labour market, including all types of e-

services(Qostal et al., 2020). They are producing and

storing massive and heterogeneous data containing

personal and professional information. Accordingly,

this data can be exploited and feed any big data

system built to understand variables and factors with

their weights and their impacts on the workplace

using different data mining techniques(Piad et al.,

2016)(Mishra et al., 2017).

Therefore, the big data concept(De Mauro et al.,

2016), as defined by De Mauro, Greco, & Grimaldi

“information assets characterized by such a high

volume, velocity and variety that they require specific

technology and analytical methods for its

transformation into value”. The fundamental

definition of big data is based on the 3Vs

characteristic:

 Volume: refers to the size of data generated.

Currently, a study conducted by IDC showed

that in 2020 the data created, captured, copied,

and consumed has reached 59 (ZB).

Furthermore, there are 69444 users creating

information in the professional world and

applying for jobs on the LinkedIn platform for

each minute. Based on information from Data

Never Sleeps 8.0.

 Variety: Generally, the data forming the big

data systems come from different sources,

which means different types and varieties of

data, generally can be grouped into three types:

structured, semi-structured, and unstructured.

Today, 90% of this data is semi-structured and

unstructured, varying from text, video, audio

metadata (data about data).

 Velocity: The speed at which data is generated

on the web varies from platform to platform,

requiring appropriate strategies to manage data

as it enters the big data system. Generally, the

applied approaches vary from batch processing

to stream processing.

Accordingly, for the labour market context, the

3Vs dimensions benefitted from the emergence of

SNS, e-learning, e-recruitment. They were giving the

base for the application of different types of analytics

in the employment context (Mishra et al., n.d.)

(Saouabi and Ezzati, 2019) (descriptive, diagnostic,

predictive and prescriptive).

2.2 Big Data for Labour Market

Intelligence System (LMIS)

Previously, collecting and treating data related to the

labour market was done through the traditional

Labour Market Information (LMI), where the main

sources of data for LMI were based essentially on

surveys and administrative records from public

employment agencies to construct the information

about the labour market and help decision-makers to

analyze and take actions on the LMP process

(Sorenson and Mas, n.d.). Although, the traditional

LMI suffered from low accuracy, outdated

classifications techniques and the lack of the ability

to turn labour market information into intelligent

decisions in a short time (Johnson, 2016). To

overcome those difficulties especially with the era of

real-time data generation with high heterogeneity and

velocity, an alternative represented with the real-time

LMI approach is developed to support data coming

from web platforms and using solutions provided

with the big data ecosystem.

Furthermore, the analytic part of the data

collected from traditional and real-time LMI is

handled by projects grouped under a big field of

research known as the Big Data for Labour Market

Intelligence System (LMIS) (Mezzanzanica and

Mercorio, 2018) (table1). Mainly, with the initiative

of projects launched by the European Centre for the

Development of Vocational Training (Cedefop)

agency since 2016 with the aim to support the use of

big data into the employment context and fellow the

undergoing process where the combination of the big

data ecosystem and the computation techniques

including machine learning algorithms is applied to

analyze labour market.

Big Data, Hadoop and Spark for Employability: Proposal Architecture

261

Table 1: Some projects supported for LMIS context.

Results

Evaluation on n-grams

extracted both from titles and

full description windows best

results was with SVM Linear:

Precision 0.910, Recall 0.909,

F1-Score 0.908, Accuracy

0.909.

The best performing

model in this case is BERT

Cloud, with an accuracy of

0.816 after 210 training steps

(epochs), followed by BERT

Local (0.8) and Simple-NN

(0.78)

Methodology

Using scarping techniques on e-recruitment platforms of the countries

involved in the project (at first stage: 6 countries) using the batch mode for the

ingestion phase (once per week over 4 months) to loads the outcomes of this

step into a data warehouse. Classify each job vacancy with respect to the

International Standard Classification of Occupations (ISCO) taxonomy. Using

supervised learning approach as machine learning techniques with three

algorithms to classify the job vacancies based on SVMs, ANN Artificial Neural

Networks and Random Forests al

orithms

The dataset was provided by UWV agency containing millions of

vacancies posts from year 2014, the training and validation steps was on 10K

vacancies and the validation was under supervision of experts from UWV.

DataOps pipeline combining Big Data and Machine Learning models to

process vacancies and resumes, Six domain experts from UWV annotated a

total of 3K sentences – around 500 sentences per person The baseline pre-

trained model used both for BERT Cloud and BERT Local for text

classification based on ISCO and ESCO taxonomy.

Description

A system developed for the

Cedefop Agency; with the aim to

create a KDD and LMIS on the

European labour market by using a

data coming from European e-

recruitment platforms

Based on the DataOps pipeline in

the context of project Werkinzicht,

author propose as LMIS system using big

data as base to create KDD for job

seekers of the Dutch-Flemish

labour market

Study/Projec

WoLMIS :(Boselli

et al., 2018)

DataOps for

Societal

Intelligence(Tamburri et

al., 2020)

The first results of the implementation of Cedefop

projects over the Union European were published in

2019 with the participation of 16 Member States

(Austria, Belgium, Czech Republic, Denmark,

Germany, Hungary, Spain, Finland, France, Italy,

Ireland, Luxembourg, the Netherlands, Poland,

Portugal, Sweden, and Slovakia). Otherwise, since

2019 the European Training Foundation ETF

launched imitative to enhance the use of LMIS within

NON-EU countries such as Morocco and Tunisia to

analysis the web labour markets and job vacancy

(OJV) portals(Vaccarino et al., n.d.). Therefore, an

initiative launched as sa result of the cooperation

between the Moroccan government and Millennium

Challenge Corporation (MCC) since 2015, via the

MCA-Morocco in October 2020, a competition to set

up a digital labour market information platform based

on artificial intelligence and big data.

2.3 Big Data Architectures for the

Employment

The solutions proposed to deal with the employability

must take into consideration the identifications

process of variables with their complexity and their

impact in the labour market, the stakeholder analysis

process(job providers, job seekers, etc.) with their

actions and reactions towards the workplace, and the

complexity of data generated related to employability

with the determination of different data sources.

Secondly, develop an advanced analytics method to

process the information collected from LMIS to

discover hidden patterns, extract relevant features,

and propose new insights to make appropriate

decisions. Accordingly, ETF proposed architectures

based on big data ecosystem, including Hadoop and

Spark solutions (

Figure 1), to guarantee scalability,

distribute tasks, and store data over Multi-node

architectures with the possibility of using advanced

analytics approaches.

The solutions proposed for LMIS have a primary

objective to overcome the previous challenges and

create the Knowledge Discovery in Data KDD related

to employability (Fayyad et al., 1996).

Figure 1: an overview of LMIS based on Hadoop & spark

solutions proposed by ETF (Vaccarino et al., n.d.).

The proposition adopted by ETF (figure 1),

divided the implementation of big data for the LMIS

into three significant steps using the Hadoop

ecosystem and apache spark.

BML 2021 - INTERNATIONAL CONFERENCE ON BIG DATA, MODELLING AND MACHINE LEARNING (BML’21)

262

2.3.1 LM Information Ingestion

The ingestion process is considered as a vital step

with the aim to integrate valuable information into the

proposed big data system (Meehan et al., n.d.). The

process is about transferring data from various data

sources (NoSQL stores, web pages using scraping

and crawling techniques, databases, SaaS data, etc.)

into the distributed storage unit of the system (data

warehouse, database, data mart or filesystem), usually

in the Hadoop ecosystem is represented with the

Hadoop Distributed File System HDFS, or throw

other propositions of NoSQL databases (e.g.

MongoDB, HBASE). Mainly, we can group the data

sources related to the labour market into four types:

 Employment websites: regroup all e-

recruitment platforms and professional social

networks like LinkedIn, where information can

be gathered for job providers and seeker’s

profiles, resumes, experiences and

responsibilities.

 Education and training platforms: represented

by educational platforms, including e-learning

and MOOCs like Udemy, Coursera, GitHub,

Slack and Sourceforge. Those platforms

provide valuable information of training

process, skills acquired, performance and

contributions on the share of technical

knowledge. The information collected can be a

way to measure the quality of skills within a

community.

 Social media platforms: Such as Facebook and

Twitter, many studies have shown the vital

information can be extracted, including

indicators on personality traits using the big

five model using NLP and sentimental analysis

to decide on the recruitment process (Satyaki

Sanyal et al., 2017).

 Government e-services: with the era of open

data adopted by the governments, the official

data can be gathered on the employability

context and the development of the labour

market to identify the weak points and skills

gaps to build a global strategy and LMP.

This step raises big challenges with the high

velocity and large volume of data generated,

requiring real-time approaches, high fault tolerance,

and simultaneous multi-connection from variant

sources. The ingestion process from those sources can

be made in two possible ways:

 Batch ingestion: considered as the most used

method. The data collection operation from the

sources is applied periodically within a

predefined time interval. The popularity of this

method comes from the optimization that can

be made on the speed and quality of data

transferred from the source in a controlled and

efficient manner.

 Real-Time Streaming: The data ingestion is

made in real-time from sources as soon as data

is available and recognized by the data

ingestion layer without respecting a predefined

schedule. The loading in real-time provides

valuable information to monitor variables and

indicators on critical situations where the need

for fast decisions.

In the Hadoop ecosystem, many solutions are

offered as an interface and broker between sources

and destinations, such as KAFKA, FLUME, and

SQOOP (Mccaffrey, 2020).

2.3.2 Data Preprocessing & LM

The data collected and retrieved from different

sources related to LM are usually in raw and noisy

forms, where the step of data preprocessing is about

the preparation and transformation of this data into a

suitable form with high-quality values to increase the

predictions and accuracy of LMIS and avoid the

unclean information which will lead to wrong results

“garbage in garbage out” (García et al., 2016) (He et

al., 2010). Fundamentally, the major task of this

phase is to deal with duplicated data such as

duplicated profiles on different platforms, missing

values like competence, skill and diplomas

information, treat outliers, encoding categorical

value, features selection and features extraction

before deciding to deploy a machine learning solution

(García et al., 2016):

 Noise reduction: the data retrieved from

platforms such as social networks contain

information that is usually unrelated to

professional activities such as emojis, hashtags,

URLs. That information can be considered

irrelevant for the analytics step; by removing

this data, accuracy and precision can be

improved.

 Merging process: this step aims to merge

duplicated information from different sources

such as vacancy announcements on different

platforms, profiles of seekers, and statistics

from various sources published on OJVs. This

step will give the actual situation of the needs

and offers in the workplace in terms of profiles

and skills.

 Normalization and standardization: the

objective of this step is to unify values and

information within a norm of measure,

especially with the data related to

Big Data, Hadoop and Spark for Employability: Proposal Architecture

263

employability become very large and the fields

of studies become verb y diversified, pushing

to have unified norms of skills, job description,

occupation and competence. In LMIS, there are

many norms, Occupational Information

Network (O*NET), ESCO and ISCO.

2.3.3 LM Knowledge and Analysis

The LM knowledge and analysis aims to use Machine

learning techniques on the data processing to generate

the KDD about labour market performance, help in the

talent management process, give recommendations to

stakeholders with answers about the quality of skills,

professions required, sectors needed to be improved.

Various analytics approaches target employability in

different stages using big data. Many application

beside LMIS has objective to enhance employability

such as Vocational Education and Training Systems

(TVET) (S. M. S. Azman Shah et al., n.d.), HRIS

systems(Kaygusuz et al., 2016) (Shahbaz et al.,

2019)where the objective use the people analytics

approach(Isson and Harriott, n.d.; Shrivastava et al.,

2018) to answers questions on employee behavior,

workforce trends, and performance at the job as a way

of hiring and hunting talents and predict turnover

probabilities (Tamburri et al., 2020). Mainly, text

classification, natural language processing and texts

Job Posting Analytics JPA are widely used as a way to

builds model capable to help us predict on match

capabilities to market needs (table 2), identify

valuable talent like never before, match capabilities to

market needs. In the applicative context, many

solutions proposed the application of the mapreduce

where the mapper and reducer function on hadoop

context or pyspark on the spark context can be used as

a way on the analyze of resume, and as results is the

helps in the process of LMP throw different models of

recommendations (Collaborative, user or hybrid

recommendation)(Li et al., 2017).

3 PROPOSED ARCHITECTURE

As explained in the previous sections, the objective is

to adopt architecture capable of exploiting and

processing the fast-growing data generated over

electronic platforms such as SNS, e-government

services, e-learning systems of universities,

employment platforms, and any potential platform

that could generate information related to youth

concerns. Accordingly, many initiatives and projects

are trying to create a KDD oriented to the

employability context using Big Data. Mainly, our

Table 2: Some studies and research for JPAs supported for

LMIS context.

Results

Results show an

improvement of jobs obtained

via different combination of

these parameters reaching 90%

of accuracy.

The proposed algorithm

ITFACP compared to

algorithm(ACR, UCB, HCT

and OLCR) perform better

when the number of instances

exceed 500,000 the regret reach

52% and 78% of the OLCR

and ACR algorithm.

Methodology/Sample

The introduction of homogeneous graph-based

recommendation architecture (GBR) based on graph

theory. Using a real data set (jobs and users) from

CareerBuilder.com on a sample of 500,000 jobs. Using

the Apache Spark GraphX framework. Job data is

stored in Hive databases, and SparkSQL.

A Contextual Online Learning Approach was

developed (such as the preference or other characteristics in

registration information and browse records). dissimilarit

function ℒ which can indicate the similarity of different

items contexts are similar to handle the diversification of user

personality. Using as a sample database from Work4

data(100 million instances) contains the candidates profile

information based on Facebook and Linkedin.

Study

an hybrid recommendation system

to improve the job search process,

based on the theory of graphs(

nodes/edges) by exploiting several

relevant values provided form job

seekers profiles(Shalaby et al., 2017).

An optimal algorithm was

developed to predict and give

recommendation concerning

employment in the big data

scenarios (S. Dong et al., 2017).

approach tries to combine the process adopted and

showed in Figure 2 to create KDD and the

implementation using the Hadoop ecosystem and

spark with their components:

 Hadoop Ecosystem: Includes packages and

subcomponents to manage Big Data throw all

stages of processing of data and support building

models that address youth concerns. The storage

part of data can be managed using HDFS where

can be dispatched over clusters as a way of fault

tolerance such as HBase and MongoDB. For the

analytics, parts can be handled directly throw

MapReduce as a way of data processing on large

clusters or throws other alternatives where the

functionality of Mapper and Reducer can be done

throw query language, for example, Hive and Pig.

 Spark Ecosystem: Apache Spark is an open-

source alternative to Hadoop and MapReduce,

where the computation is 100 times faster. The

use of Spark within the step of preprocessing data

repose on the RDD (Resilient Distributed

Datasets) to adapt data for the processing and

machine learning to develop recommendations,

predictive insights and personalized, where Spark

BML 2021 - INTERNATIONAL CONFERENCE ON BIG DATA, MODELLING AND MACHINE LEARNING (BML’21)

264

Figure 2: the proposed architecture based on Hadoop & spark and the KDD model (Fayyad et al., 1996).

is equipped MLlib library includes a package

aimed to address these concerns. Combining the

Hadoop ecosystem and Spark on the management

and processing layer will improve and enhance

data processing and provide real-time analysis of

youth data. The model (Figure 2) will help to

identify hidden patterns and potential solutions in

many areas, such as education and

employment(He et al., 2010), offering new

opportunities to improve policy design for the

youth generation and LMP.

4 CONCLUSIONS

The application of big data in different contexts has

proven to be efficient and effective; for that, our

research tries to find different approaches to apply big

data and machine learning within the employability

context for the youth generation. Therefore, the

ongoing process and research to create LMIS based

on big data based on different solutions and visions

can be a road map for further research about creating

an intelligent system based on interaction with

different stakeholders and data coming from different

sources.

REFERENCES

Boselli, R., Cesarini, M., Marrara, S., Mercorio, F.,

Mezzanzanica, M., Pasi, G., Viviani, M., 2018.

WoLMIS: a labor market intelligence system for

classifying web job vacancies. J Intell Inf Syst 51, 477–

502. https://doi.org/10.1007/s10844-017-0488-x

Caliendo, M., 2016. Youth Unemployment and Active

Labor Market Policies in Europe 30.

De Mauro, A., Greco, M., Grimaldi, M., 2016. A formal

definition of Big Data based on its essential features.

Library Review 65, 122–135.

https://doi.org/10.1108/LR-06-2015-0061

Fayyad, U., Piatetsky-Shapiro, G., Smyth, P., 1996. The

KDD process for extracting useful knowledge from

volumes of data. Commun. ACM 39, 27–34.

https://doi.org/10.1145/240455.240464

García, S., Ramírez-Gallego, S., Luengo, J., Benítez, J.M.,

Herrera, F., 2016. Big data preprocessing: methods and

prospects. Big Data Anal 1, 9.

https://doi.org/10.1186/s41044-016-0014-0

He, Q., Tan, Q., Ma, X., Shi, Z., 2010. The High-Activity

Parallel Implementation of Data Preprocessing Based

on MapReduce, in: Yu, J., Greco, S., Lingras, P., Wang,

G., Skowron, A. (Eds.), Rough Set and Knowledge

Technology, Lecture Notes in Computer Science.

Springer Berlin Heidelberg, Berlin, Heidelberg, pp.

646–654. https://doi.org/10.1007/978-3-642-16248-

0_88

Isson, J.P., Harriott, J.S., n.d. People Analytics in the Era of

Big Data 15.

Johnson, E., 2016. Can Big Data Save Labor Market

Information Systems? RTI Press.

https://doi.org/10.3768/rtipress.2016.pb.0010.1608

Kaygusuz, İ., Akgemci, T., Yilmaz, A., 2016. The Impact

of HRIS Usage on Organizational Efficiency and

Employee Performance: A Research in Industrial and

Banking Sector in Ankara and Istanbul Cities. IJoBM

IV. https://doi.org/10.20472/BM.2016.4.4.002

Li, Z., Lin, Y., Zhang, X., 2017. Hybrid employment

recommendation algorithm based on Spark. J. Phys.:

Conf. Ser. 887, 012045. https://doi.org/10.1088/1742-

6596/887/1/012045

Mccaffrey, P., 2020. Overview of big data tools: Hadoop,

Spark, and Kafka, in: An Introduction to Healthcare

Informatics. Elsevier, pp. 291–305.

https://doi.org/10.1016/B978-0-12-814915-7.00020-X

Meehan, J., Tatbul, N., Aslantas, C., Zdonik, S., n.d. Data

Ingestion for the Connected World 11.

Mezzanzanica, M., Mercorio, F., 2018. Big Data Enables

Labor Market Intelligence, in: Sakr, S., Zomaya, A.

(Eds.), Encyclopedia of Big Data Technologies.

Springer International Publishing, Cham, pp. 1–11.

TVET

Big Data, Hadoop and Spark for Employability: Proposal Architecture

265

Mishra, S.N., Lama, D.R., Pal, Y., n.d. Human Resource

Predictive Analytics (HRPA) for HR Management in

Organizations 5, 3.

Mishra, T., MewarUniversity, Chittorgarh – 312901,

Rajasthan, India, Kumar, D., Department of Computer

Science, G. J. University, Hisar – 125001, Haryana,

India, Gupta, S., Guru Nanak Institute of Management,

West Punjabi Bagh – 110026, Delhi, India, 2017.

Students’ Performance and Employability Prediction

through Data Mining: A Survey. Indian Journal of

Science and Technology 10, 1–6.

https://doi.org/10.17485/ijst/2017/v10i24/110791

Piad, K.C., Dumlao, M., Ballera, M.A., Ambat, S.C., 2016.

Predicting IT employability using data mining

techniques, in: 2016 Third International Conference on

Digital Information Processing, Data Mining, and

Wireless Communications (DIPDMWC). Presented at

the 2016 Third International Conference on Digital

Information Processing, Data Mining, and Wireless

Communications (DIPDMWC), pp. 26–30.

https://doi.org/10.1109/DIPDMWC.2016.7529358

Qostal, A., Moumen, A., Lakhrissi, Y., 2020. Systematic

Literature Review on Big Data and Data Analytics for

Employment of Youth People: Challenges and

Opportunities:, in: Proceedings of the 2nd International

Conference on Advanced Technologies for Humanity.

Presented at the International Conference on Advanced

Technologies for Humanity, SCITEPRESS - Science

and Technology Publications, Rabat, Morocco, pp.

179–185. https://doi.org/10.5220/0010432201790185

Ryder, G., Director-General, I., 2020. World Employment

and Social Outlook - Trends 2020. World Employment

and Social Outlook 108.

S. Dong, Z. Lei, P. Zhou, K. Bian, G. Liu, 2017. Job and

Candidate Recommendation with Big Data Support: A

Contextual Online Learning Approach, in:

GLOBECOM 2017 - 2017 IEEE Global

Communications Conference. Presented at the

GLOBECOM 2017 - 2017 IEEE Global

Communications Conference, pp. 1–7.

https://doi.org/10.1109/GLOCOM.2017.8255006

S. M. S. Azman Shah, , H. N. Haron2, , M. N. Mahrin, n.d.

The Trend of Big Data in Workforce Frameworks and

Occupational Standards towards an Educational

Intelligent Economy. JTET.

https://doi.org/10.30880/jtet. 2021.13.01.019

Saouabi, M., Ezzati, A., 2019. Proposition of an

employability prediction system using data mining

techniques in a big data environment. International

Journal of Mathematics and Computer Science 14,

411–424.

Satyaki Sanyal, Souvik Hazra, Neelanjan Ghosh,

Soumyashree Adhikary, 2017. Resume Parser with

Natural Language Processing.

https://doi.org/10.13140/RG.2.2.11709.05607

Shahbaz, U., Beheshti, A., Nobari, S., Qu, Q., Paik, H.-Y.,

Mahdavi, M., 2019. iRecruit: Towards Automating the

Recruitment Process, in: Lam, H.-P., Mistry, S. (Eds.),

Service Research and Innovation, Lecture Notes in

Business Information Processing. Springer

International Publishing, Cham, pp. 139–152.

https://doi.org/10.1007/978-3-030-32242-7_11

Shalaby, W., Alaila, B., Korayem, M., Pournajaf, L.,

Aljadda, K., Quinn, S., Zadrozny, W., 2017. Help me

find a job: A graph-based approach for job

recommendation at scale, in: Nie J.-Y., Obradovic Z.,

Suzumura T., Ghosh R., Nambiar R., Wang C., Zang

H., Baeza-Yates R., Baeza-Yates R., Hu X., Kepner J.,

Cuzzocrea A., Tang J., Toyoda M. (Eds.), Proc. - IEEE

Int. Conf. Big Data, Big Data. Institute of Electrical and

Electronics Engineers Inc., pp. 1544–1553.

https://doi.org/10.1109/BigData.2017.8258088

Shrivastava, S., Nagdev, K., Rajesh, A., 2018. Redefining

HR using people analytics: the case of Google. HRMID

26, 3–6. https://doi.org/10.1108/HRMID-06-2017-

0112

Sorenson, K., Mas, J.-M., n.d. A Roadmap for the

Development of Labor Market Information Systems

2.2.3 LMIS in Select Countries 74.

Tamburri, D.A., Heuvel, W.-J.V.D., Garriga, M., 2020.

DataOps for Societal Intelligence: a Data Pipeline for

Labor Market Skills Extraction and Matching, in: 2020

IEEE 21st International Conference on Information

Reuse and Integration for Data Science (IRI). Presented

at the 2020 IEEE 21st International Conference on

Information Reuse and Integration for Data Science

(IRI), IEEE, Las Vegas, NV, USA, pp. 391–394.

https://doi.org/10.1109/IRI49571.2020.00063

Vaccarino, A., Mezzanzanica, M., Mercorio, F., Castel-

Branco, E., n.d. Methodological overview and

Analytics insights on Tunisian Web Labour Market 45.

BML 2021 - INTERNATIONAL CONFERENCE ON BIG DATA, MODELLING AND MACHINE LEARNING (BML’21)

266