Using Machine Learning for Labour Market Intelligence

Boselli, Roberto; Cesarini, Mirko; Mercorio, Fabio; Mezzanzanica, Mario

doi:10.1007/978-3-319-71273-4_27

Cited by 29 publications

(20 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The results from such methods can be directly interpreted and presented as findings. Some research approaches use the vector space model [59], [107], e.q., numerical datasets created from the text process, as input for further quantitative methods (see Figure 7) appropriate for numerical data.…”

Section: Rq3: How Can Labor Market Data Be Analyzed? (Methods)mentioning

confidence: 99%

“…5) Ontology: Capturing labor market trends from the content of online job advertisements is an important aspect. In this paper, we identified primary studies which utilize existing labor market ontologies such as ESCO [101], [107] or O*NET [49] to map job titles from collected advertisements with existing titles in these ontologies. However, a new challenge is the creation of a new ontology which describes the whole digital labor content.…”

Section: ) Adoption By Policy Makers (Stakeholder: Policymentioning

confidence: 99%

See 1 more Smart Citation

Extracting Knowledge From On-Line Sources for Software Engineering Labor Market: A Mapping Study

et al. 2019

View full text Add to dashboard Cite

Software engineering is a continuously evolving sector and the demands of the related labor market result in a wide variety of job openings, ranging from developers to customer service positions. Thus, there is a need to continuously monitor labor market trends using data and analytics. Both employers and employees can benefit by capturing emerging trends which can facilitate continuous learning and training in new technologies, support of better matching between a job offer and the ideal candidate and expertise detection. To fulfill these needs, the results of labor market analytics need to reach the stakeholders timely and accurately. However, often delays occur, which stem from time-consuming approaches based on collecting data from traditional sources, such as questionnaires or interviews. Recently, researchers started leveraging content from digital sources, which are easily accessed and contain a wealth of information. This paper presents the results of a Systematic Mapping Study on digital sources that can be used to address the data analytics needs of the labor market. It provides a multifaceted categorization of the issues involved in the analysis of digital sources of the software engineering labor market. It aims to identify digital labor market sources for data retrieval which are appropriate for employers and employees analytics. Additionally, it aims to connect different skill types, needs and goals of labor market with the utilization of digital sources and data analysis methods. In total 86 primary studies were selected and each one was evaluated and classified aiming to identify the: (a) digital sources that are used for labor market analytics; (b) type of skills they examine; (c) methods which are used to utilize the raw digital content; (d) goals for which every primary study is conducted; (e) beneficiaries (stakeholder) of the results; and (f) time trends for all the above.

show abstract

Section: Rq3: How Can Labor Market Data Be Analyzed? (Methods)mentioning

confidence: 99%

Section: ) Adoption By Policy Makers (Stakeholder: Policymentioning

confidence: 99%

Extracting Knowledge From On-Line Sources for Software Engineering Labor Market: A Mapping Study

et al. 2019

View full text Add to dashboard Cite

show abstract

“…Then, we match the job titles in our dataset which also exist in the ESCO taxonomy, and use the matched job titles for the experiment. While there is another publicly available job title taxonomy ISCO 3 , ESCO is an extended version of ISCO and has a much larger volume. Thus, we adopted ESCO and created the ground truth label using the job title defined by ESCO.…”

Section: Datasetmentioning

confidence: 99%

“…We compare JAMES against a exhaustive list of ten baseline models relevant to the JTM task: KNN-based [38], Word2Vec-based [3], DeepCarotene [32], Node2Vec [14], GloVe [27], NEO [13], WoLMIS [2], BERT-based [30], Job2Vec [35], and Universal Sentence Encoder (USE) [5]. The details of baselines and our implementation settings are described in the Appendix A.…”

Section: Set-upmentioning

confidence: 99%

JAMES: Job Title Mapping with Multi-Aspect Embeddings and Reasoning

Yamashita¹,

Shen²,

Ekhtiari³

et al. 2022

Preprint

View full text Add to dashboard Cite

One of the most essential tasks needed for various downstream tasks in career analytics (e.g., career trajectory analysis, job mobility prediction, and job recommendation) is Job Title Mapping (JTM), where the goal is to map user-created (noisy and non-standard) job titles to predefined and standard job titles. However, solving JTM is domain-specific and non-trivial due to its inherent challenges:(1) user-created job titles are messy, (2) different job titles often overlap their job requirements, (3) job transition trajectories are inconsistent, and (4) the number of job titles in real world applications is large-scale. Toward this JTM problem, in this work, we propose a novel solution, named as JAMES, that constructs three unique embeddings of a target job title: topological, semantic, and syntactic embeddings, together with multi-aspect co-attention. In addition, we employ logical reasoning representations to collaboratively estimate similarities between messy job titles and standard job titles in the reasoning space. We conduct comprehensive experiments against ten competing models on the large-scale real-world dataset with more than 350,000 job titles. Our results show that JAMES significantly outperforms the best baseline by 10.06% in Precision@10 and by 17.52% in NDCG@10, respectively. Recently [20] built a 30,000 job title taxonomy on LinkedIn for a job understanding task. However, little is known about how the JTM task with the aforementioned challenges can be solved.Proposed Ideas. Toward these challenges, in this paper, we propose JAMES (Job title mApping with Multi-aspect Embeddings and rea Soning) to solve the JTM task. We use a large-scale and real-world career dataset with more than 350,000 job titles that a

show abstract

“…Several classifiers were extensively evaluated for this task: linear SVM, RBF kernel SVM, fully connected neural networks, and convolutional neural networks working on word embedding (the interested reader can refer to Boselli [ ] for further details on the evaluated classifiers, the optimization activities, and the parameters tuned). The latter evaluation was performed on a scenario similar to the one described in this paper but focusing on English vacancies.…”

Section: Web Scraping and Classificationmentioning

confidence: 99%

Skills in demand for ICT and statistical occupations: Evidence from web‐based job vacancies

Lovaglio

Cesarini

Mercorio

et al. 2018

Statistical Analysis

Self Cite

View full text Add to dashboard Cite

Online job portals collecting web vacancies have become important media for job demand and supply matching. They also represent a growing research area for the application of analytical methods to study the labour market using innovative data sources. This paper analyses Italian web job vacancies scraped from several types of Italian web job portals between June and September 2015. After describing how the occupations associated with each web vacancy (classification up to level 4) were identified and the related skills retrieved in texts using mixed supervised and unsupervised text mining approaches, we focused on job vacancies related to ICT and statistical positions. The principal aim of this paper is to describe these jobs in terms of the required skills that have emerged in the labour market from a demand perspective and to identify those skills that best distinguish statisticians from other ICT occupations. Hence, several machine learning techniques were used to assess those skills that best distinguish occupation codes from other job groups. After quality control and removal of duplications, the scraping collected more than 110,000 job advertisements: nearly 6,200 were classified as ICT or statistical positions (largely dominated by software developers). The data indicate that high‐level statisticians have superior and heterogeneous professional backgrounds, linked to theoretical statistics, where analytic skills are more relevant than computing skills. Many soft and management‐oriented skills were also called for, which are missing among lower level statisticians, who are restricted to more technical jobs oriented towards general computing and informatics.

show abstract

Using Machine Learning for Labour Market Intelligence

Cited by 29 publications

References 23 publications

Extracting Knowledge From On-Line Sources for Software Engineering Labor Market: A Mapping Study

Extracting Knowledge From On-Line Sources for Software Engineering Labor Market: A Mapping Study

JAMES: Job Title Mapping with Multi-Aspect Embeddings and Reasoning

Skills in demand for ICT and statistical occupations: Evidence from web‐based job vacancies

Contact Info

Product

Resources

About