Files in this item



application/pdfSPIRIN-DISSERTATION-2016.pdf (12MB)
(no description provided)PDF


Title:Optimizing search user interfaces and interactions within professional social networks
Author(s):Spirin, Nikita Valeryevich
Director of Research:Karahalios, Karrie G.
Doctoral Committee Chair(s):Karahalios, Karrie G.
Doctoral Committee Member(s):Zhai, ChengXiang; Han, Jiawei; Tunkelang, Daniel
Department / Program:Computer Science
Discipline:Computer Science
Degree Granting Institution:University of Illinois at Urbana-Champaign
Subject(s):Professional Social Network
Online Social Network
Search User Interface
Search Snippet
Information Extraction
Job Search
People Search
User Study
Interview Study
Need Elicitation
User Profiling
Social Search
Query Log Analysis
E-commerce Search
Structured Search
Query Understanding
Log Mining
Search User Interface Optimization
User Behavior
User Modeling
Human-Computer Interaction
Information Retrieval
Machine Learning
Graph Search
Entity Search
Search Ranking
Search Engine Result Page
Learning to Rank
Information Filtering
Expert Search
Abstract:Professional social networks (PSNs) play the key role in the online social media ecosystem, generate hundreds of terabytes of new data per day, and connect millions of people. To help users cope with the scale and influx of new information, PSNs provide search functionality. However, most of the search engines within PSNs today still provide only keyword queries, basic faceted search capabilities, and uninformative query-biased snippets overlooking the structured and interlinked nature of PSN entities. This results in siloed information, inefficient results presentation, and suboptimal search user experience (UX). In this thesis, we reconsider and comprehensively study input, control, and presentation elements of the search user interface (SUI) to enable more effective and efficient search within PSNs. Specifically, we demonstrate that: (1) named entity queries (NEQs) and structured queries (SQs) complement each other helping PSN users search for people and explore the PSN social graph beyond the first degree; (2) relevance-aware filtering saves users' efforts when they sort jobs, status updates, and people by an attribute value rather than by relevance; (3) extended informative structured snippets increase job search effectiveness and efficiency by leveraging human intelligence and exposing the most critical information about jobs right on a search engine result page (SERP); and (4) non-redundant delta snippets, which different from traditional query-biased snippets show on a SERP information relevant but complementary to the query, are more favored by users performing entity (e.g. people) search, lead to faster task completion times and better search outcomes. Thus, by modeling the structured and interlinked nature of PSN entities, we can optimize the query-refine-view interaction loop, facilitate serendipitous network exploration, and increase search utility. We believe that the insights, algorithms, and recommendations presented in this thesis will serve the next generation designers of SUIs within and beyond PSNs and shape the (structured) search landscape of the future.
Issue Date:2016-07-14
Rights Information:Copyright 2016 by Nikita Valeryevich Spirin. All rights reserved.
Date Available in IDEALS:2016-11-10
Date Deposited:2016-08

This item appears in the following Collection(s)

Item Statistics