An Ensemble Machine Learning Approach for Classifying Job Positions

Abstract

Machine learning is one of the promising research areas in computer science, with numerous applications in automated detection of meaningful data patterns. Several data-centric studies were conducted on evaluating competencies, detecting similar jobs and predicting salaries of various job positions. However, the hazy distinction between closely related job positions requires powerful predictive algorithms. The present study proposed an ensemble approach for accurate classification of various job positions. Accordingly, different machine learning algorithms were applied on 955 instances obtained from Glassdoor using web scraping. Furthermore, the present study classify various job positions based on average salary and other correlated explanatory variables that cover many aspects of job activities on the internet. The study result revealed the superior performance of heterogeneous ensembles in terms of precision and accuracy. The proposed data-centric approach produce strong models for researchers, recruiters, and candidates to assigned job positions and its competencies.