Skip to main content
All CollectionsData MethodologyPostings
Lightcast Occupations Taxonomy (LOT) Classification Methodology
Lightcast Occupations Taxonomy (LOT) Classification Methodology

The LOT classifier - How it works

Updated over a week ago

The Lightcast Occupations Taxonomy (LOT) Classifier predicts LOT Specialized Occupations using a curated machine learning model approach.

The classifier uses a job posting's title, description and regional information (i.e. origin country and language) in order to classify the content to one of Lightcast's Specialized Occupations. Taxonomists curate parts of the model to target the specific keywords, skills, and terminology of our Specialized Occupations. The remaining portion of the model learns from a corpus of raw job postings data. This ensures that our occupation classifier is picking up on all the skills, relevant certifications, representative roles and responsibilities, and other industry or occupation specific terminology in order to classify the posting correctly. Common words, company boilerplates, and benefits are removed prior to classification.

Once classified to a specialized occupation within Lightcast's LOT Taxonomy, the posting can then be mapped to a national taxonomy code of interest (e.g., O*NET, UKSOC, NOC, etc.). The LOT Classifier produces the derived occupation names and codes for the specialized occupation structure. This structure is comprised of the Career Area, Occupation Group, Occupation and Specialized Occupation. From this data, relevant national taxonomy codes can be further derived.

Did this answer your question?