Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Standard Occupation Classifier -- A Natural Language Processing Approach

Sidharth Rony, Jack Patman

TL;DR

The paper tackles automated SOC-code assignment for noisy online job advertisements by deploying a BERT-based ensemble that fuses job titles, descriptions, and extracted skills. It systematically investigates truncation strategies, preprocessing, and hierarchical versus flat classification, reporting that a combined title-description-skills ensemble yields the strongest performance, with up to about 61% accuracy at the most granular tier and 72% at a coarser tier. The work demonstrates the viability of transformer-based approaches for cross-language SOC mapping (UK ONS SOC and US O*NET SOC) using real-world data and highlights practical challenges such as long text, data sparsity at deep taxonomy levels, and taxonomy revisions. Collectively, the findings offer a scalable, data-driven method for real-time labor market analysis and occupation-specific demand insights from job postings, with implications for workforce planning and policy evaluation.

Abstract

Standard Occupational Classifiers (SOC) are systems used to categorize and classify different types of jobs and occupations based on their similarities in terms of job duties, skills, and qualifications. Integrating these facets with Big Data from job advertisement offers the prospect to investigate labour demand that is specific to various occupations. This project investigates the use of recent developments in natural language processing to construct a classifier capable of assigning an occupation code to a given job advertisement. We develop various classifiers for both UK ONS SOC and US O*NET SOC, using different Language Models. We find that an ensemble model, which combines Google BERT and a Neural Network classifier while considering job title, description, and skills, achieved the highest prediction accuracy. Specifically, the ensemble model exhibited a classification accuracy of up to 61% for the lower (or fourth) tier of SOC, and 72% for the third tier of SOC. This model could provide up to date, accurate information on the evolution of the labour market using job advertisements.

Standard Occupation Classifier -- A Natural Language Processing Approach

TL;DR

The paper tackles automated SOC-code assignment for noisy online job advertisements by deploying a BERT-based ensemble that fuses job titles, descriptions, and extracted skills. It systematically investigates truncation strategies, preprocessing, and hierarchical versus flat classification, reporting that a combined title-description-skills ensemble yields the strongest performance, with up to about 61% accuracy at the most granular tier and 72% at a coarser tier. The work demonstrates the viability of transformer-based approaches for cross-language SOC mapping (UK ONS SOC and US O*NET SOC) using real-world data and highlights practical challenges such as long text, data sparsity at deep taxonomy levels, and taxonomy revisions. Collectively, the findings offer a scalable, data-driven method for real-time labor market analysis and occupation-specific demand insights from job postings, with implications for workforce planning and policy evaluation.

Abstract

Standard Occupational Classifiers (SOC) are systems used to categorize and classify different types of jobs and occupations based on their similarities in terms of job duties, skills, and qualifications. Integrating these facets with Big Data from job advertisement offers the prospect to investigate labour demand that is specific to various occupations. This project investigates the use of recent developments in natural language processing to construct a classifier capable of assigning an occupation code to a given job advertisement. We develop various classifiers for both UK ONS SOC and US O*NET SOC, using different Language Models. We find that an ensemble model, which combines Google BERT and a Neural Network classifier while considering job title, description, and skills, achieved the highest prediction accuracy. Specifically, the ensemble model exhibited a classification accuracy of up to 61% for the lower (or fourth) tier of SOC, and 72% for the third tier of SOC. This model could provide up to date, accurate information on the evolution of the labour market using job advertisements.

Paper Structure

This paper contains 40 sections, 3 equations, 14 figures, 13 tables.

Table of Contents

  1. Introduction
  2. Background
  3. Standard Occupational Classification
  4. Literature Review
  5. Language Models
  6. Large Document Analysis
  7. General Overview
  8. Data
  9. Data Collection
  10. Data Preparation
  11. Data Summary
  12. Methodology
  13. Preprocessing
  14. Train, Test Split
  15. Model
  16. ...and 25 more sections

Figures (14)

  • Figure 2.1: UK ONS SOC Classifier structure for an aircraft maintenance occupation.
  • Figure 2.2: Subset of the ONS SOC 2020 classification demonstrating the taxonomy.
  • Figure 2.3: US O*NET SOC Classifier structure for a survey research occupation.
  • Figure 2.4: The SOC2020 taxonomy - demonstrating the significant stratification of the final label classifications. Different colours are representative of different 'major group' occupational codes.
  • Figure 2.5: Workflow of Occupation Classification
  • ...and 9 more figures