Publication Type

Conference Proceeding Article

Version

publishedVersion

Publication Date

9-2018

Abstract

Career track represents a vertical career pathway, where one can gradually move up to take up higher job appointments when relevant skills are acquired. Understanding the propensity of career movements in an evolving job market can enable timely career guidance to job seekers and working professionals. To this end, we harvest career trajectories from online professional network (OPN). Our focus lies on obtaining a macro view on career movements at the track granularity. Specifically, we propose a semi-supervised career track labelling framework to automatically assign career tracks for large set of jobs. To contextually label jobs, we collect example jobs with career track labels identified by human resource specialists and domain experts in Singapore. An intuitive idea is to learn the labelling knowledge from the example jobs and then apply to jobs in OPN. Unfortunately, such small amount of labeled jobs presents a great challenge in our attempt to accurately recover career tracks for plentiful unlabelled jobs. We thus address the issue by resorting to semi-supervised learning methods. This research not only reduces the human annotation efforts in maintaining the career track knowledge databases over time across different geographical regions, but also facilitates data science study on career movements. Extensive experiments are conducted to demonstrate the labelling accuracy as well as to gain insights upon obtained career track labels.

Keywords

Label Propagation, Career Movements Analysis, Career Track Labelling

Discipline

Databases and Information Systems

Research Areas

Data Science and Engineering

Publication

Workshop on Data Science for Human Capital Management (DSHCM2018), Dublin, Ireland, 2018, September 14

First Page

1

Last Page

16

Publisher

DSHCM

City or Country

Dublin

Copyright Owner and License

Authors

Download

Share

COinS