Model-based approach to tag current and previous work experience to the best corresponding O*NET-SOC Taxonomy occupation labels. Using this standard classification would make our data easier to map to US Department of Labor and 3rd party data, including for employment statistic and job role skills requirements.
The model-based approach to the tagging itself will also allow for a greater number of job experience to be tagged (increasing fill rates), and for variables such as the company and industry of the employee to be factored into job role predictions (increasing accuracy).