Cross-Platform Expertise Characterization for Question Routing

Master thesis (2017)

Authors

A. Rawat Electrical Engineering, Mathematics and Computer Science

Contributors

A. Bozzon (supervisor 1)
J. Yang (supervisor 1)
Faculty
Electrical Engineering, Mathematics and Computer Science
Expertise Cross-Platform Question Routing
More Info
expand_more

Published Date

31-10-2017

Language

English

Reuse Rights

Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.

Faculty

Electrical Engineering, Mathematics and Computer Science

Abstract

Community Question-Answer(CQA) services are online portals with a community of people that share similar interests, and post question-answers for solving the difficulties they have. These services have huge social impact and have become a place where knowledge is created and distributed. Despite the advances in Question Routing and Recommendation Systems, CQA still struggles in handling large number of unanswered questions and suffer with the quality of answers provided in general. An effective solution to tackle this problem is to consider the expertise of users when routing a question to its potential answerers. There are many existing methods for measuring expertise but they tend to favour more active users rather than expert users. To address this problem,
we look into cross-platform behaviour of a user
and incorporate his cross-platform expertise. We provide insights into performance of question routing with this method.

We perform an extensive literature survey about the CQA system, question routing, recommendation systems and types of expertise, to map a users' cross-platform behavior with a certain type of expertise. We propose a methodology to match users across multiple platforms, with which we measure expertise according to mapped user behaviour on those platforms. Finally we estimate the reliability of expertise measurement when applied to question routing. We study and discuss the effect of parameters such as division of data into training and test set sizes and use factorization machine for recommendation system.