A Network Science and Document Similarity based Hybrid Job Recommendation System
Name
Maksym Sukhorukov
Abstract
Job recommendation systems mainly use different sources of data in order to give the better content for the end user. Developing the well-performing system requires complex hybrid approaches of representing similarity based on the content of job postings and resumes as well as interactions between them. We develop an efficient hybrid network-based job recommendation system which uses Personalized PageRank algorithm in order to rank vacancies for the users based on the similarity between resumes and job posts as textual documents, along with previous interactions of users with vacancies. Our approach achieved the recall of 50% and generated more applies for the jobs during the online A/B test than previous algorithms.
Graduation Thesis language
English
Graduation Thesis type
Master - Computer Science
Supervisor(s)
Rajesh Sharma
Defence year
2018