Browse Prior Art Database

System and Method for Finding Candidates for a Job Using Social Behavioral Analysis

IP.com Disclosure Number: IPCOM000235823D
Publication Date: 2014-Mar-26
Document File: 5 page(s) / 102K

Publishing Venue

The IP.com Prior Art Database

Abstract

Selecting the best candidate and the future best performer for a job is a very important problem. Companies receive thousands of applications for each job posting and assessing them manually is nearly impossible. In this disclosure, we propose to use the Social Profile of candidates as an indicator of their job performance. Our method can also handle the different behavioral traits that are required for different jobs. This describes a three-stage system for finding and ranking candidates for a job by behavioral analysis from social media data, where : Stage 1: The system initially learns a model based on a set of social behavioral predictors of the top performers for that job . Stage 2: For a new candidate(s), the system inputs their social behavioral patterns/predictors (used in (1)) and outputs their predicted performance based on the learned model and ranks them according to the prediction. Stage 3: The system iteratively refines the learned model of (1) based on the observed/actual performance of candidates hired as the outcome of (2).

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 36% of the total text.

Page 01 of 5

System and Method for Finding Candidates for a Job Using Social Behavioral Analysis

Introduction:


¥ Selecting the best candidate and the future best performer for a job is a very important problem:

¥ Resume-based: Matching candidate's educational qualifications and experience to job requirements is increasingly becoming automated. E.g. Prospect [A. Singh et al., Prospect: A system for screening candidates for recruitment, CIKM 2010].

¥ Social and behavioral aspects of the candidates play a major role in their later performance in the organization [Ostroff et al., The relationship between satisfaction, attitudes, and performance: An organizational level analysis, Journal of Applied Psychology, 1992]


¥ Recruiters are increasingly inspecting the social profiles of applied candidates to filter out those whose professionalism is of concern [ref: Eurocom Worldwide Survey 2012: http://www.eurocompr.com/prfitem.asp?id=14921 Retrieved: 28 Feb, 2013]


¥ Companies receive thousands of applications for each job posting:


¥ Manual thorough and comprehensive interview/selection process is difficult and doesn't scale.


¥ "The Eurocom Worldwide survey reveals that nearly half (49%) of technology executives say that their firm will increase their expenditure on social media in the next 12 months"


¥ Study: Social Profile indicative of job performance

¥

http://blogs.wsj.com/digits/2012/02/22/could-your-facebook-profile-predict-j

  ob-performance/
¥ Studies only generic traits; not specific to any job.
¥ Different jobs require different behavioral traits. Example,
¥ Public Relations Manager might be:


¥ Very active in social media, posting and responding in a number of conversations


¥ Very large social network and associated with a large number of communities


¥ Organizing a number of events to promote the company / some cause


¥ While, a Researcher may be:


¥ Less active in social media, posting only to very selected discussions


¥ Social network may be limited to only former classmates and colleagues


¥ Following some technical blogs/authors, but not necessarily

1


Page 02 of 5

organizing events


¥ A three-stage system for finding and ranking candidates for a job by behavioral analysis from social media data, where :
Stage 1: The system initially learns a model based on a set of social behavioral predictors of the top performers for that job :


¥ For each top performer, the system access one or more sources of their social interaction to collect relevant data


¥ System finds and learns to weigh different social behavioral patterns that are indicative of the job performance (these social aspects are detailed later in the deck)


¥ The system also identifies the peer group of the candidate based on the amount of interaction and uses the social behavioral predictors of people in the peer group, weighed by the closeness of the peer to the candidate (as measured by the amount of interaction) also while learning the model.

Stage 2: For a new candidat...