Job Classification Based on Skills and Qualifications Using Natural Language Processing and Ensemble Learning Methods

Hafiza Oktasia Nasution; Dian Ramadhani; Mida Aprilina Tarigan; Prima Andreas; Dewita Suryati Ningsih; Arwinence Pramadewi

doi:10.25299/itjrd.2025.25550

Authors

Hafiza Oktasia Nasution Department of Business Economics, University of Riau
Dian Ramadhani Department of Business Economics, University of Riau
Mida Aprilina Tarigan Department of Business Economics, University of Riau
Prima Andreas Department of Business Economics, University of Riau
Dewita Suryati Ningsih Department of Business Economics, University of Riau
Arwinence Pramadewi Department of Business Economics, University of Riau

DOI:

https://doi.org/10.25299/itjrd.2025.25550

Keywords:

Job Classification, Natural Language Processing, XGBoost, Ensemble Learning, Workforce Analytics

Abstract

This study proposes a job classification framework using Natural Language Processing (NLP) and Ensemble Learning to classify job roles based on required skills and qualifications. A large-scale open-source dataset containing 1.048.576 job postings was utilized, with attributes such as job title, qualifications, skills, company profile, and role. Only relevant attributes were used: skills and qualifications as input features, and role as the target label. Data were filtered to focus on three major job roles—Management, IT, and Digital—resulting in 489,651 relevant entries. Skills were extracted and standardized using GROK AI before feature transformation with MultiLabelBinarizer for one-hot encoding. The XGBoost algorithm was applied for classification under multiple data split configurations (70:15:15, 80:10:10, 70:30, 80:20, 90:10) with random_state=42 and multi-class log loss evaluation. Results showed that the 90:10 configuration achieved the highest accuracy (74.18%), followed by 80:20 with 68.44%. This research demonstrates that ensemble learning effectively handles high-dimensional categorical job data and provides a foundation for automated job classification systems and labor market analysis.

Downloads

Download data is not yet available.

References

[1] V. S. Litvinenko, “Digital economy as a factor in the technological development of the mineral sector,” Nat. Resour. Res., vol. 29, no. 3, pp. 1521–1541, 2020.

[2] M. J. Sousa and D. Wilks, “Sustainable skills for the world of work in the digital age,” Syst. Res. Behav. Sci., vol. 35, no. 4, pp. 399–405, 2018.

[3] A. Bhola, K. Halder, A. Prasad, and M. Y. Kan, “Retrieving Skills from Job Descriptions: A Language Model Based Extreme Multi-label Classification Framework,” COLING 2020 - 28th Int. Conf. Comput. Linguist. Proc. Conf., pp. 5832–5842, 2020, doi: 10.18653/v1/2020.coling-main.513.

[4] D. Zare, L. Fernandez-Sanz, V. Pospelova, and I. López-Baldominos, “NLP and Text Mining for Enriching IT Professional Skills Frameworks,” Appl. Sci., vol. 15, no. 17, pp. 1–20, 2025, doi: 10.3390/app15179634.

[5] E. Senger, M. Zhang, R. van der Goot, and B. Plank, “Deep Learning-based Computational Job Market Analysis: A Survey on Skill Extraction and Classification from Job Postings,” NLP4HR 2024 - 1st Work. Nat. Lang. Process. Hum. Resour. Proc. Work., no. Nlp4hr, pp. 1–15, 2024.

[6] A. Famili, W.-M. Shen, R. Weber, and E. Simoudis, “Data preprocessing and intelligent data analysis,” Intell. data Anal., vol. 1, no. 1–4, pp. 3–23, 1997.

[7] S. García, J. Luengo, and F. Herrera, Data preprocessing in data mining, vol. 72. Springer, 2015.

[8] M. E. de Carvalho Souza and L. Weigang, “Grok, gemini, chatgpt and deepseek: Comparison and applications in conversational artificial intelligence,” Intel. Artif., vol. 2, no. 1, 2025.

[9] M. M. Abushaega, O. Y. Moshebah, A. Hamzi, and S. Y. Alghamdi, “Multi-objective sustainability optimization in modern supply chain networks: A hybrid approach with federated learning and graph neural networks,” Alexandria Eng. J., vol. 115, pp. 585–602, 2025.

[10] W. Gao, Z. Ding, J. Lu, and Y. Wan, “Low-carbon information quality dimensions and random forest algorithm evaluation model in digital marketing,” Sci. Rep., vol. 14, no. 1, p. 22416, 2024.

[11] T. Chen et al., “Xgboost: extreme gradient boosting,” R Packag. version 0.4-2, vol. 1, no. 4, pp. 1–4, 2015.

[12] T. Chen, T. He, M. Benesty, and V. Khotilovich, “Package ‘xgboost,’” R version, vol. 90, no. 1–66, p. 40, 2019.

[13] J. Dong, Y. Chen, B. Yao, X. Zhang, and N. Zeng, “A neural network boosting regression model based on XGBoost,” Appl. Soft Comput., vol. 125, p. 109067, 2022.

[14] X. Zhang, C. Yan, C. Gao, B. A. Malin, and Y. Chen, “Predicting missing values in medical data via XGBoost regression,” J. Healthc. informatics Res., vol. 4, no. 4, pp. 383–394, 2020.

[15] N. Dawson, M.-A. Rizoiu, B. Johnston, and M.-A. Williams, “Predicting skill shortages in labor markets: A machine learning approach,” in 2020 IEEE International Conference on Big Data (Big Data), 2020, pp. 3052–3061.

[16] A. A. Syahputra and R. E. Saputro, “Application of the XGBoost Model with Hyperparameter Tuning for Industry Classification for Job Applicants,” Sinkron, vol. 8, no. 3, pp. 1920–1931, 2024, doi: 10.33395/sinkron.v8i3.13840.

[17] J. Chen, Y. Chen, Y. Liao, J. Mu, G. Li, and J. Li, “Job Recruitment Analysis based on Xgboost Decision Tree,” Int. J. Soc. Sci. Public Adm., vol. 2, no. 3, pp. 392–396, 2024, doi: 10.62051/ijsspa.v2n3.56.

[18] M. A. Khan, M. O. Khan, H. Noureen, M. S. Khan, and M. Fawad, “A Hybrid Transformer and CNN-Based Approach for Classifying Mental Health Disorders from Social Media Data,” 2025.

[19] M. Tajrian, A. Rahman, M. A. Kabir, and M. R. Islam, “Analysis of child development facts and myths using text mining techniques and classification models,” Heliyon, vol. 10, no. 17, 2024.

[20] A. I. A. Osman, A. N. Ahmed, M. F. Chow, Y. F. Huang, and A. El-Shafie, “Extreme gradient boosting (Xgboost) model to predict the groundwater levels in Selangor Malaysia,” Ain Shams Eng. J., vol. 12, no. 2, pp. 1545–1556, 2021.