Working paper Open Access

Identifying AI talents among LinkedIn members, A machine learning approach

Thomas Roca


Dublin Core Export

<?xml version='1.0' encoding='utf-8'?>
<oai_dc:dc xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:oai_dc="http://www.openarchives.org/OAI/2.0/oai_dc/" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" xsi:schemaLocation="http://www.openarchives.org/OAI/2.0/oai_dc/ http://www.openarchives.org/OAI/2.0/oai_dc.xsd">
  <dc:creator>Thomas Roca</dc:creator>
  <dc:date>2019-04-23</dc:date>
  <dc:description>How to identify specific profiles among the hundred of millions gathered in LinkedIn? LinkedIn Economic Graph thrives on skills,
around 50 thousand of them are listed by LinkedIn and constitute one of the main signals to identify professions or trends. Artificial Intelligence (AI) skills, for example, can be used to identify the diffusion of AI in industries [16]. But the noise can be loud around skills for which the demand is high. Some users may add "trendy" skills on their profiles without having work experience or training related to them. On the other hand, some people may work in the broad AI ecosystem (e.g. AI recruiters, AI sales representatives, etc.), without being the AI practitioners we are looking for. Searching for keywords in profiles' sections can lead to mis-identification of certain profiles, especially for those related to a field rather than an occupation. This is the
case for Artificial Intelligence. In this paper, we propose a machine learning approach to identify such profiles, and suggest
to train a binary text-classifier using job offers posted on the platform rather than actual profiles.
We suggest this approach allows to avoid manually labeling the training dataset, granted the assumption that job profiles posted by recruiters are more "ideal-typical" or simply provide a more consistent triptych "job title, job description, associated skills" than the ones that can be found among member's profiles.</dc:description>
  <dc:identifier>https://zenodo.org/record/3240963</dc:identifier>
  <dc:identifier>10.5281/zenodo.3240963</dc:identifier>
  <dc:identifier>oai:zenodo.org:3240963</dc:identifier>
  <dc:language>eng</dc:language>
  <dc:relation>doi:10.5281/zenodo.2649207</dc:relation>
  <dc:relation>url:https://zenodo.org/communities/dfp17</dc:relation>
  <dc:rights>info:eu-repo/semantics/openAccess</dc:rights>
  <dc:rights>http://creativecommons.org/licenses/by/4.0/legalcode</dc:rights>
  <dc:subject>Artificial Intelligence</dc:subject>
  <dc:subject>Skills</dc:subject>
  <dc:subject>Machine learning</dc:subject>
  <dc:subject>Natural Language Processing</dc:subject>
  <dc:subject>Big Data</dc:subject>
  <dc:title>Identifying AI talents among LinkedIn members, A machine learning approach</dc:title>
  <dc:type>info:eu-repo/semantics/workingPaper</dc:type>
  <dc:type>publication-workingpaper</dc:type>
</oai_dc:dc>
210
155
views
downloads
All versions This version
Views 210148
Downloads 15594
Data volume 551.5 MB336.9 MB
Unique views 187137
Unique downloads 14188

Share

Cite as