Developing and Evaluating a Query Recommendation Feature to Assist Users with Online Information Seeking & Retrieval With graduate students: Karl Gyllstrom,

Slides:



Advertisements
Similar presentations
The Robert Gordon University School of Engineering Dr. Mohamed Amish
Advertisements

The Cost of Authoring with a Knowledge Layer Judy Kay and Lichao Li School of Information Technologies The University of Sydney, Australia.
1 Evaluation Rong Jin. 2 Evaluation  Evaluation is key to building effective and efficient search engines usually carried out in controlled experiments.
Optimizing search engines using clickthrough data
UCLA : GSE&IS : Department of Information StudiesJF : 276lec1.ppt : 5/2/2015 : 1 I N F S I N F O R M A T I O N R E T R I E V A L S Y S T E M S Week.
Exercising these ideas  You have a description of each item in a small collection. (30 web sites)  Assume we are looking for information about boxers,
Overview of Collaborative Information Retrieval (CIR) at FIRE 2012 Debasis Ganguly, Johannes Leveling, Gareth Jones School of Computing, CNGL, Dublin City.
Dialogue – Driven Intranet Search Suma Adindla School of Computer Science & Electronic Engineering 8th LANGUAGE & COMPUTATION DAY 2009.
Evaluating Search Engine
© Tefko Saracevic, Rutgers University1 digital libraries and human information behavior Tefko Saracevic, Ph.D. School of Communication, Information and.
1 CS 430: Information Discovery Lecture 22 Non-Textual Materials 2.
1 CS 430 / INFO 430 Information Retrieval Lecture 8 Query Refinement: Relevance Feedback Information Filtering.
Modern Information Retrieval
Model Personalization (1) : Data Fusion Improve frame and answer (of persistent query) generation through Data Fusion (local fusion on personal and topical.
INFO 624 Week 3 Retrieval System Evaluation
Retrieval Evaluation. Brief Review Evaluation of implementations in computer science often is in terms of time and space complexity. With large document.
© Tefko Saracevic, Rutgers University1 digital libraries and human information behavior Tefko Saracevic, Ph.D. School of Communication, Information and.
Reference Collections: Task Characteristics. TREC Collection Text REtrieval Conference (TREC) –sponsored by NIST and DARPA (1992-?) Comparing approaches.
Design of metadata surrogates in search result interfaces of learning object repositories: Linear versus clustered metadata design Panos Balatsoukas Anne.
Advance Information Retrieval Topics Hassan Bashiri.
Lessons Learned from Information Retrieval Chris Buckley Sabir Research
An investigation of query expansion terms Gheorghe Muresan Rutgers University, School of Communication, Information and Library Science 4 Huntington St.,
Language Modeling Frameworks for Information Retrieval John Lafferty School of Computer Science Carnegie Mellon University.
Personalizing the Digital Library Experience Nicholas J. Belkin, Jacek Gwizdka, Xiangmin Zhang SCILS, Rutgers University
© Tefko Saracevic, Rutgers University1 digital libraries and human information behavior Tefko Saracevic, Ph.D. School of Communication, Information and.
Usability and Evaluation Dov Te’eni. Figure ‎ 7-2: Attitudes, use, performance and satisfaction AttitudesUsePerformance Satisfaction Perceived usability.
PBT and the Web The Feasibility of Teaching English in Rural Russia Leon Gipson Web Developer and Systems Analyst ECU Ludmila Tataru Ludmila Tataru Chair.
This chapter is extracted from Sommerville’s slides. Text book chapter
류 현 정류 현 정 Human Computer Interaction Introducing evaluation.
Personalization in Local Search Personalization of Content Ranking in the Context of Local Search Philip O’Brien, Xiao Luo, Tony Abou-Assaleh, Weizheng.
Search Engines and Information Retrieval Chapter 1.
Dafna Hardbattle, Ken Fisher & Peter Chalk London Metropolitan University International Blended Learning Conference University of Hertfordshire,
1 The BT Digital Library A case study in intelligent content management Paul Warren
Evaluation Experiments and Experience from the Perspective of Interactive Information Retrieval Ross Wilkinson Mingfang Wu ICT Centre CSIRO, Australia.
Philosophy of IR Evaluation Ellen Voorhees. NIST Evaluation: How well does system meet information need? System evaluation: how good are document rankings?
Understanding and Predicting Graded Search Satisfaction Tang Yuk Yu 1.
Personalizing Information Search: Understanding Users and their Interests Diane Kelly School of Information & Library Science University of North Carolina.
1 Formal Models for Expert Finding on DBLP Bibliography Data Presented by: Hongbo Deng Co-worked with: Irwin King and Michael R. Lyu Department of Computer.
David W. Klein Helen A. Schartz AERA National Conference Vancouver, B.C., Canada April 16, 2012 Instructional Strategies to Improve Informed Consent in.
A Simple Unsupervised Query Categorizer for Web Search Engines Prashant Ullegaddi and Vasudeva Varma Search and Information Extraction Lab Language Technologies.
Xiaoying Gao Computer Science Victoria University of Wellington Intelligent Agents COMP 423.
UOS 1 Ontology Based Personalized Search Zhang Tao The University of Seoul.
Exploring Online Social Activities for Adaptive Search Personalization CIKM’10 Advisor : Jia Ling, Koh Speaker : SHENG HONG, CHUNG.
ZLOT Prototype Assessment John Carlo Bertot Associate Professor School of Information Studies Florida State University.
1 Information Retrieval Acknowledgements: Dr Mounia Lalmas (QMW) Dr Joemon Jose (Glasgow)
Implicit Acquisition of Context for Personalization of Information Retrieval Systems Chang Liu, Nicholas J. Belkin School of Communication and Information.
RCDL Conference, Petrozavodsk, Russia Context-Based Retrieval in Digital Libraries: Approach and Technological Framework Kurt Sandkuhl, Alexander Smirnov,
Effects of Popularity and Quality on the Usage of Query Suggestions during Information Search Can users be induced to take bad query suggestions because.
TOPIC CENTRIC QUERY ROUTING Research Methods (CS689) 11/21/00 By Anupam Khanal.
Dataware’s Document Clustering and Query-By-Example Toolkits John Munson Dataware Technologies 1999 BRS User Group Conference.
Personalized Course Navigation Based on Grey Relational Analysis Han-Ming Lee, Chi-Chun Huang, Tzu- Ting Kao (Dept. of Computer Science and Information.
Evaluation of a Visualization System for Information Retrieval at the Front and the Back End Gregory B. Newby Sch of Information and Lib. Science U. of.
Personalized Interaction With Semantic Information Portals Eric Schwarzkopf DFKI
Chapter 8 Evaluating Search Engine. Evaluation n Evaluation is key to building effective and efficient search engines  Measurement usually carried out.
Measuring How Good Your Search Engine Is. *. Information System Evaluation l Before 1993 evaluations were done using a few small, well-known corpora of.
Information Retrieval
1 Chapter 12 Configuration management This chapter is extracted from Sommerville’s slides. Text book chapter 29 1.
ASSOCIATIVE BROWSING Evaluating 1 Jinyoung Kim / W. Bruce Croft / David Smith for Personal Information.
Evaluation INST 734 Module 5 Doug Oard. Agenda Evaluation fundamentals Test collections: evaluating sets Test collections: evaluating rankings Interleaving.
The Loquacious ( 愛說話 ) User: A Document-Independent Source of Terms for Query Expansion Diane Kelly et al. University of North Carolina at Chapel Hill.
Evaluation of Information Retrieval Systems Xiangming Mu.
Identifying “Best Bet” Web Search Results by Mining Past User Behavior Author: Eugene Agichtein, Zijian Zheng (Microsoft Research) Source: KDD2006 Reporter:
DTC Quantitative Methods Measurement II: Questionnaire Design and Scale Construction Friday 27 th January 2012.
ASSOCIATIVE BROWSING Evaluating 1 Jin Y. Kim / W. Bruce Croft / David Smith by Simulation.
Navigation Aided Retrieval Shashank Pandit & Christopher Olston Carnegie Mellon & Yahoo.
Bringing Order to the Web : Automatically Categorizing Search Results Advisor : Dr. Hsu Graduate : Keng-Wei Chang Author : Hao Chen Susan Dumais.
WHIM- Spring ‘10 By:-Enza Desai. What is HCIR? Study of IR techniques that brings human intelligence into search process. Coined by Gary Marchionini.
IR Theory: Evaluation Methods
Retrieval Performance Evaluation - Measures
Measuring Learning During Search: Differences in Interactions, Eye-Gaze, and Semantic Similarity to Expert Knowledge Florian Groß Mai
Presentation transcript:

Developing and Evaluating a Query Recommendation Feature to Assist Users with Online Information Seeking & Retrieval With graduate students: Karl Gyllstrom, Earl Bailey Diane Kelly, Assistant Professor University of North Carolina at Chapel Hill

Background  Query formulation is one of the most important and difficult aspects of information seeking  Users often need to enter multiple queries to investigate different aspects of their information needs  Some techniques have been developed to assist users with query formulation and reformulation:  Term Suggestion  Query Recommendation  However, there are problems associated with each of these techniques … ALISE Conference | January 23, 2009 | Denver, CO

Problems  Term Suggestion  Works via relevance feedback (often times ‘pseudo’ relevance feedback is used which makes assumptions about the goodness of the initial query)  Users don’t have the additional cognitive resources to engage in explicit feedback (‘form’ is awkward)  Users are too lazy to provide feedback – principle of least effort (‘form’ is cumbersome)  Terms are not presented in context so it may be hard for users to understand how they can help ALISE Conference | January 23, 2009 | Denver, CO

Problems  Query Suggestion  It is hard to determine the similarity of previous queries to one another (and to the current query)  Sparsity problem: assumes a set of queries that are similar to the current query exists ALISE Conference | January 23, 2009 | Denver, CO

Our Approach  User Query: dog law enforcement SUGGESTED TERMS Canine Legal Charge Train Drug Traffic Police Search Officer Dog law enforcement canine Canine legal drug traffic Dog law police enforcement drug Dog law police drug search SUGGESTED QUERIES ALISE Conference | January 23, 2009 | Denver, CO

Studies  Study I (System/Algorithm Evaluation no Users)  Identify and evaluate techniques for identify terms from corpus given a query  Identify and evaluate techniques for using these terms to create effective and semantically meaningful queries  Studies II-IV (Interactive Evaluation with Users)  Evaluate automatic query suggestion techniques, including  Comparison with term suggestions  Comparison with user-generated suggestions  Investigation of effects of topic difficulty and familiarity  Compare ‘remote’ study mode with laboratory study mode ALISE Conference | January 23, 2009 | Denver, CO

Study I: Some Questions  How do we identify the best terms from the corpus given the user’s query?  How do we select the best terms from those generated?  In what order do we combine terms?  How do we incorporate the initial query?  How long should the recommended queries be?  How many queries do we suggest?  Our Solution  Implemented Tan, et al.’s (2007) clustering method for selecting terms (language modeling framework)  TREC-style evaluation using a test collection ALISE Conference | January 23, 2009 | Denver, CO

Studies II-IV: Common Elements  Two interfaces: Query Suggestion and Term SuggestionQuery Suggestion  Each subject completed two search topics with each interface  Task: Find and save documents relevant to the information described in the topic  Up to 15 minutes to search per topic  Twenty search topics in total sorted into four difficulty levels : Easy, Medium, Moderate, Difficult  Each subject completed one topic from each level  Rotation and counter-balancing …  Subjects searched a closed corpus of over 1 million newspaper articles (AP, NYT and XN) ALISE Conference | January 23, 2009 | Denver, CO

Studies II-IV: Common Elements  Several outcome measures:  Use of suggestions (System Log)  Performance (Retrieval Results and Docs Saved)  Mean Average Precision (Baseline Relevance Assessments)  Interactive Precision and Recall (Integrate BRA with User RA)  Discounted Cumulated Gain (User RA)  Perceived Effectiveness and Satisfaction (Exit Questionnaire)  Preference (Exit Questionnaire)  Qualitative Feedback (Exit Questionnaire) ALISE Conference | January 23, 2009 | Denver, CO

Studies II-IV: Common Elements  And a few more independent variables:  Topic Difficulty (Pre-determined Level)  Subject’s Topic Knowledge (Pre-topic Questionnaire)  Subject’s Experienced Difficulty (Exit Questionnaire) ALISE Conference | January 23, 2009 | Denver, CO

Studies II-IV: Common Procedures START END Pre-Topic Questionnaire [Repeat for 2 Systems] Exit Questionnaire Consent Subject Searches [Repeat for 2 Topics] Demographic Questionnaire Search Experience Questionnaire ALISE Conference | January 23, 2009 | Denver, CO

Studies II-IV: Differences  Study II (n=43)  Subjects completed this study remotely  Study III (n=25)  Eye-tracking data collected from first 12 subjects  Study IV (n=22)  Additional qualitative data collection via stimulated recall for two searches (one per system)  Study III and IV  Variation in Source of Suggestions: Half received system- generated suggestions (same as Study II) and half received user- generated suggestions (extracted from Study II subjects) ALISE Conference | January 23, 2009 | Denver, CO

Preliminary Results  Use ALISE Conference | January 23, 2009 | Denver, CO

Preliminary Results  Use and Source of Suggestions ALISE Conference | January 23, 2009 | Denver, CO

Preliminary Results  Use & Topic ALISE Conference | January 23, 2009 | Denver, CO

Preliminary Results  Perceived Effectiveness and Satisfaction  For 7 of the 11 Exit Questionnaire items, query suggestion was rated higher than term suggestion. These items concerned:  ‘Cognitive Assistance’ (e.g., helped me think more about the topic and understand its different aspects)  Satisfaction  Term suggestion was rated higher with respect to  Modification  Ease of Use  There were few differences in ratings of system-generated suggestions and user-generated suggestions ALISE Conference | January 23, 2009 | Denver, CO

Preliminary Results ALISE Conference | January 23, 2009 | Denver, CO  Preference

Next Steps  Continue data analysis …  Impact of topic difficulty and knowledge  Eye-tracking data  ‘Typing’ of suggestions  Temporal/Stage Analysis ALISE Conference | January 23, 2009 | Denver, CO

BACK ALISE Conference | January 23, 2009 | Denver, CO