Employing Natural Language Processing Techniques for Online Job Vacancies Classification

Varelas, George; Lagios, Dimitris; Ntouroukis, Spyros; Zervas, Panagiotis; Parsons, Kenia; Tzimas, Giannis

doi:10.1007/978-3-031-08341-9_27

Cited by 5 publications

(5 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…They observed that while both models perform similarly in classifying job employers, GBDT is more effective than SVM in identifying job employers that were wrongly computed. Varelas G. et al [28] also used SVM, random forests, KNN, SGD (stochastic gradient descent), and MLP neural network classifiers as voting algorithms for the classification of job postings according to the ISCO Occupation Codes.…”

Section: Previous Workmentioning

confidence: 99%

“…For the critical output layer, we have employed the softmax activation function, aligning with established best practices for multi-class classification tasks. The number of units in this layer corresponds to the previously calculated unique class count (28), ensuring comprehensive coverage of class predictions.…”

Section: Models Architecture and Trainingmentioning

confidence: 99%

“…Tokenizer and Model: The tokenizer, obtained from the "BERT-base-uncased" pretrained model, is responsible for processing text input. The model, also pretrained on "BERT-base-uncased", is tailored for sequence classification and has a number of labels equal to the count of unique class labels (28) in the training data. This architecture is particularly well-suited for text classification tasks [44].…”

Section: Models Architecture and Trainingmentioning

confidence: 99%

“…Data Preparation: For training and evaluation, we created two datasets, one for training and one for evaluation. These datasets include text inputs and their corresponding category labels (28), which are encoded as numerical values. The tokenizer is applied to tokenize and preprocess the text data, and a maximum sequence length of 256 tokens is imposed.…”

Section: Models Architecture and Trainingmentioning

confidence: 99%

See 3 more Smart Citations

Deep Learning Approaches for Big Data-Driven Metadata Extraction in Online Job Postings

Skondras,

Zotos,

Lagios

et al. 2023

Information

Self Cite

View full text Add to dashboard Cite

This article presents a study on the multi-class classification of job postings using machine learning algorithms. With the growth of online job platforms, there has been an influx of labor market data. Machine learning, particularly NLP, is increasingly used to analyze and classify job postings. However, the effectiveness of these algorithms largely hinges on the quality and volume of the training data. In our study, we propose a multi-class classification methodology for job postings, drawing on AI models such as text-davinci-003 and the quantized versions of Falcon 7b (Falcon), Wizardlm 7B (Wizardlm), and Vicuna 7B (Vicuna) to generate synthetic datasets. These synthetic data are employed in two use-case scenarios: (a) exclusively as training datasets composed of synthetic job postings (situations where no real data is available) and (b) as an augmentation method to bolster underrepresented job title categories. To evaluate our proposed method, we relied on two well-established approaches: the feedforward neural network (FFNN) and the BERT model. Both the use cases and training methods were assessed against a genuine job posting dataset to gauge classification accuracy. Our experiments substantiated the benefits of using synthetic data to enhance job posting classification. In the first scenario, the models’ performance matched, and occasionally exceeded, that of the real data. In the second scenario, the augmented classes consistently outperformed in most instances. This research confirms that AI-generated datasets can enhance the efficacy of NLP algorithms, especially in the domain of multi-class classification job postings. While data augmentation can boost model generalization, its impact varies. It is especially beneficial for simpler models like FNN. BERT, due to its context-aware architecture, also benefits from augmentation but sees limited improvement. Selecting the right type and amount of augmentation is essential.

show abstract

Section: Previous Workmentioning

confidence: 99%

Section: Models Architecture and Trainingmentioning

confidence: 99%

Section: Models Architecture and Trainingmentioning

confidence: 99%

Section: Models Architecture and Trainingmentioning

confidence: 99%

See 2 more Smart Citations

Deep Learning Approaches for Big Data-Driven Metadata Extraction in Online Job Postings

Skondras,

Zotos,

Lagios

et al. 2023

Information

Self Cite

View full text Add to dashboard Cite

show abstract

“…Extracting data structures with text mining has illustrated the relevant job skills in the medical informatics area and the skills required for computer vision and NLP specialists have been identified using the web crawling technique [12]. NLP methods are already used in the scientific context for the automatic extraction of relevant information from texts and are also efficient methods to summarize data [13], [14]. These studies have some limitations, such as the manual checking of labeling skills only by researchers and not validated by professionals in their fields, not explaining systematic semantic analysis, and in-depth analysis of job profiles in the IT field have not been found much.…”

Section: Introductionmentioning

confidence: 99%

Information Technology Job Profile Using Average-Linkage Hierarchical Clustering Analysis

Siswipraptini,

Warnars,

Ramadhan

et al. 2023

IEEE Access

View full text Add to dashboard Cite

The growth in Information Technology (IT) jobs is predicted to reach 15 percent between 2021 and 2031. The growth of IT jobs has resulted in a remarkable change in all infrastructure, such as information, skills, and domains covered in IT job profiles. Unfortunately, job roles and skills in this field remain undefined. The gap between the supply and demand needs in the IT workforce must be filled immediately with an appropriate strategy. To fulfill industry needs, an in-depth analysis of IT job profiles is important. Therefore, it is important for educational programs to identify the competencies needed by the industry to update their output. This study aims to identify the job profiles required for IT job specialists by analyzing real-world job posts published online to identify hidden meanings from a textual database. A systematic semantic methodology was proposed using an average-linkage hierarchical clustering analysis. It resembles a tree structure technique to discover relevant phrases, relationships, and hidden meanings through semantic analysis. Occurrences of the most frequent words and phrases were extracted to reveal the domain knowledge of each IT job cluster. The result is a systematic semantic analysis of the IT job profile comprising the programming language, specialized type, duty, database, tools, and frameworks. The justification for each job profile was validated by 10 IT professionals from various private and government companies in Indonesia through Focus Group Discussions (FGD).INDEX TERMS Information technology job profile, skills, Average-Linkage Hierarchical Clustering analysis, most frequent word, most frequent phrase.

show abstract

Job Title Prediction and Recommendation System for IT Professionals

Rahman,

Habiba,

Roy

et al. 2023

2023 5th International Conference on Sustainable Technologies for Industry 5.0 (STI)

View full text Add to dashboard Cite

Employing Natural Language Processing Techniques for Online Job Vacancies Classification

Cited by 5 publications

References 12 publications

Deep Learning Approaches for Big Data-Driven Metadata Extraction in Online Job Postings

Deep Learning Approaches for Big Data-Driven Metadata Extraction in Online Job Postings

Information Technology Job Profile Using Average-Linkage Hierarchical Clustering Analysis

Job Title Prediction and Recommendation System for IT Professionals

Contact Info

Product

Resources

About