FREE ELECTRONIC LIBRARY - Dissertations, online materials

Pages:   || 2 | 3 | 4 | 5 |   ...   | 19 |

«Learning Implicit User Interest Hierarchy for Web Personalization by Hyoung-rae Kim A dissertation submitted to Florida Institute of Technology in ...»

-- [ Page 1 ] --

Learning Implicit User Interest Hierarchy for

Web Personalization


Hyoung-rae Kim

A dissertation submitted to

Florida Institute of Technology

in partial fulfillment of the requirements

for the degree of

Doctor of Philosophy


Computer Science

Melbourne, Florida

May 2005


© Copyright 2005 Hyoung-rae Kim

All Rights Reserved

The author grants permission to make single copies ___________________________

Learning Implicit User Interest Hierarchy for Web Personalization a dissertation by Hyoung-rae Kim Approved as to style and content Philip K. Chan, Ph.D.

Associate Professor, Computer Sciences Dissertation Advisor Debasis Mitra, Ph.D.

Associate Professor, Computer Sciences Marius-Calin Silaghi, Ph.D.

Assistant Professor, Computer Sciences Alan C. Leonard, Ph.D.

Professor, Biological Sciences William D. Shoaff, Ph.D.

Associate Professor, Computer Sciences Department Head Abstract Learning Implicit User Interest Hierarchy for Web Personalization by Hyoung-rae Kim Dissertation Advisor: Philip K. Chan, Ph.D.

Most web search engines are designed to serve all users in a general way, without considering the interests of individual users. In contrast, personalized web search engines incorporate an individual user's interests when choosing relevant web pages to return. In order to provide a more robust context for personalization, a user interest hierarchy (UIH) is presented. The UIH extracts a continuum of general to specific user interests from web pages and generates a uniquely personalized order to search results.

This dissertation consists of five main parts. First, a divisive hierarchical clustering (DHC) algorithm is proposed to group words (topics) into a hierarchy where more general interests are represented by a larger set of words. Second, a variable-length phrase-finding (VPF) algorithm that finds meaningful phrases from a web page is introduced. Third, two new desirable properties that a correlation function should satisfy are proposed. These properties will help understand the general characteristics of a correlation function and help choose or devise correct correlation functions for an application domain. Fourth, methods are examined that (re)rank the results from a search engine depending on user interests based on the contents of a web page and the UIH. Fifth, previously studied implicit

–  –  –

new indicators are examined in more detail as well.

Experimental results indicate that the personalized ranking methods presented in this study, when used with a popular search engine, can yield more relevant web pages for individual users. The precision/recall analysis showed that our weighted term scoring function could provide more accurate ranking than Google on average.

–  –  –

List of Figures ----------------------------------------------------------------------------------------- x List of Tables -----------------------------------------------------------------------------------------xii Acknowledgements ---------------------------------------------------------------------------------xiii

1. Introduction --------------------------------------------------------------------------------------- 1

1.1. Motivation ----------------------------------------------------------------------------------- 2

1.2. Problem Statement ------------------------------------------------------------------------- 4

1.3. Approach ------------------------------------------------------------------------------------ 7

1.4. Key Contributions -------------------------------------------------------------------------- 9

1.5. Dissertation Organization ----------------------------------------------------------------10

2. Learning Implicit User Interest Hierarchy for Context in Personalization --------------12

2.1. User Interest Hierarchy -------------------------------------------------------------------14

2.2. Building User Interest Hierarchy--------------------------------------------------------17 2.2.1. Algorithm --------------------------------------------------------------------------17 2.2.2. Correlation Functions ------------------------------------------------------------23 AEMI--------------------------------------------------------------------------23 AEMI-SP ---------------------------------------------------------------------24 Other Correlation Functions------------------------------------------------26 2.2.3. Threshold-finding Methods -----------------------------------------------------27 Valley -------------------------------------------------------------------------28 MaxChildren -----------------------------------------------------------------29 Other Threshold-finding Methods-----------------------------------------30 2.2.4. Window Size and Minimum Size of a Cluster--------------------------------30

2.3. Experiments --------------------------------------------------------------------------------31 2.3.1. Experimental Data and Procedures---------------------------------------------31 2.3.2. Evaluation Criteria ---------------------------------------------------------------32

2.4. Results and Analysis ----------------------------------------------------------------------34 2.4.1. Building UIH with Only Words as Features ----------------------------------34 v Correlation Functions -------------------------------------------------------34 Threshold-finding Method--------------------------------------------------35 Window Size -----------------------------------------------------------------35 2.4.2. Building UIH with Words and Phrases as Features--------------------------37

2.5. Summary------------------------------------------------------------------------------------39

3. Identifying Variable-Length Meaningful Phrases with Correlation Functions ---------40

3.1. Variable-length Phrases ------------------------------------------------------------------42 3.1.1. VPF Algorithm--------------------------------------------------------------------42 3.1.2. Correlation Functions ------------------------------------------------------------48

3.2. Experiments --------------------------------------------------------------------------------50 3.2.1. Experimental Data and Procedures---------------------------------------------50 3.2.2. Evaluation Criteria ---------------------------------------------------------------52

3.3. Results and Analysis ----------------------------------------------------------------------54 3.3.1. With-pruning vs. Without-pruning ---------------------------------------------54 3.3.2. Analysis with Exact Match------------------------------------------------------54 Top 10 Methods--------------------------------------------------------------55 Comparing with Human Subjects -----------------------------------------57 3.3.3. Analysis with Simple Match ----------------------------------------------------58

3.4. Summary------------------------------------------------------------------------------------59

4. Analysis of Desirable Properties of Correlation Functions between Two Events ------61

4.1. Desirable Properties of a Correlation Function----------------------------------------63 4.1.1. Enhancing Property 1 ------------------------------------------------------------64 4.1.2. Additional Desirable Properties ------------------------------------------------64

4.2. Experiments --------------------------------------------------------------------------------68 4.2.1. Experimental Data and Procedures---------------------------------------------68 4.2.2. Evaluation Criteria ---------------------------------------------------------------69

4.3. Results and Analysis ----------------------------------------------------------------------71 4.3.1. Comparing Properties: Old verses New ---------------------------------------71 4.3.2. Comparison Based upon Property 1 -------------------------------------------73 4.3.3. Comparison Based upon Property 6 -------------------------------------------73 4.3.4. Normalized Results – Property 7 -----------------------------------------------74 vi

4.4. Summary------------------------------------------------------------------------------------75

5. Personalized Ranking of Search Results with Implicitly Learned User Interest Hierarchies---------------------------------------------------------------------------------------78

5.1. Personalized Results ----------------------------------------------------------------------80

5.2. Approach -----------------------------------------------------------------------------------81 5.2.1. Four Characteristics of a Matching Term -------------------------------------82 Level/Depth of a UIH Node------------------------------------------------83 Length of a Term ------------------------------------------------------------84 Frequency of a Term --------------------------------------------------------84 Emphasis of a Term ---------------------------------------------------------85 5.2.2. Scoring a Term--------------------------------------------------------------------85 Uniform Scoring -------------------------------------------------------------85 Weighted Scoring------------------------------------------------------------86 5.2.3. Scoring a Page --------------------------------------------------------------------87 5.2.4. Incorporating Public Page Score -----------------------------------------------88

5.3. Experiments --------------------------------------------------------------------------------89

5.4. Results and Analysis ----------------------------------------------------------------------91 5.4.1. Interesting Web Page-------------------------------------------------------------92 Top Link Analysis -----------------------------------------------------------92 Statistical Significance ------------------------------------------------------93 Precision/Recall Analysis---------------------------------------------------94 Varying Personal Weight ---------------------------------------------------95 5.4.2. Potentially Interesting Web Page-----------------------------------------------98 Top Link Analysis -----------------------------------------------------------98 Statistical Significance ------------------------------------------------------98 Precision/Recall Analysis---------------------------------------------------99 Varying Personal Weight ------------------------------------------------- 102

5.5. Summary---------------------------------------------------------------------------------- 102

6. Implicit Indicators for Interesting Web Pages --------------------------------------------- 105

6.1. Implicit Interest Indicators ------------------------------------------------------------- 107 6.1.1. Complete Duration ------------------------------------------------------------- 107 vii 6.1.2. Active Window Duration ------------------------------------------------------ 107 6.1.3. Look At It Duration ------------------------------------------------------------ 108 6.1.4. Distance of Mouse Movement ------------------------------------------------ 108 6.1.5. Number of Mouse Clicks ------------------------------------------------------ 109 6.1.6. Distance of Scrollbar Movement --------------------------------------------- 110 6.1.7. Number of Scrollbar Clicks --------------------------------------------------- 110 6.1.8. Number of Key UP and Down------------------------------------------------ 111 6.1.9. Size of Highlighting Text------------------------------------------------------ 111 6.1.10. Other Indicators----------------------------------------------------------------- 112

6.2. Detecting Face Orientation------------------------------------------------------------- 112 6.2.1. Detecting Three Dots----------------------------------------------------------- 113 6.2.2. Learning Face Orientation----------------------------------------------------- 117 Input/Output Parameters -------------------------------------------------- 117 Learning Algorithm ------------------------------------------------------- 118

6.3. Experiments ------------------------------------------------------------------------------ 119 6.3.1. Experimental Data and Procedures------------------------------------------- 119 6.3.2. Evaluation Criteria ------------------------------------------------------------- 120

6.4. Results and Analysis -------------------------------------------------------------------- 121 6.4.1. Visits with Maximum Duration----------------------------------------------- 121 6.4.2. All Visits------------------------------------------------------------------------- 122 6.4.3. Other Indicators----------------------------------------------------------------- 125

6.5. Summary---------------------------------------------------------------------------------- 128

7. Related Work ---------------------------------------------------------------------------------- 130

7.1. Web Information Retrieval------------------------------------------------------------- 130 7.1.1. Basics of a WIR System ------------------------------------------------------- 132 Lexical Analysis ----------------------------------------------------------- 132 Phrase ----------------------------------------------------------------------- 133 7.1.2. Clustering Web Contents ------------------------------------------------------ 136 7.1.3. Predicting Navigation ---------------------------------------------------------- 138 7.1.4. Personalized Contents---------------------------------------------------------- 139 7.1.5. Assisting Personal Information ----------------------------------------------- 142 viii 7.1.6. Implicit Detection of User’s Characteristics -------------------------------- 142

7.2. User Modeling --------------------------------------------------------------------------- 144 7.2.1. Adaptive Hypermedia---------------------------------------------------------- 145 7.2.2. Human Behavior Based User Model----------------------------------------- 146 7.2.3. Contents Based User Model--------------------------------------------------- 149 7.2.4. Hybrid Way Based User Model ---------------------------------------------- 151 7.2.5. Explicit/Implicit Way of Building a User Model -------------------------- 151

Pages:   || 2 | 3 | 4 | 5 |   ...   | 19 |

Similar works:

«RIVER HYDROAND MORPHODYNAMICS: RESTORATION, MODELING, AND UNCERTAINTY by Ari Joseph Posner _ A Dissertation Submitted to the Faculty of the DEPARTMENT OF HYDROLOGY AND WATER RESOURCES In Partial Fulfillment of the Requirements For the Degree of DOCTOR OF PHILOSOPHY WITH A MAJOR IN HYDROLOGY In the Graduate College THE UNIVERSITY OF ARIZONA THE UNIVERSITY OF ARIZONA GRADUATE COLLEGE As members of the Dissertation Committee, we certify that we have read the dissertation prepared by Ari J. Posner...»

«Enactments of Change: Becoming Textually Active at Youthline NZ by Ailsa Janet Haxell BA MHSc Submitted in fulfilment of the requirements for the degree of Doctor of Philosophy Deakin University October, 2012 DEAKIN UNIVERSITY ACCESS TO THESIS A I am the author of the thesis entitled Enactments of change: Becoming textually active at Youthline NZ submitted for the degree of Doctor of Philosophy (Education) This thesis may be made available for consultation, loan and limited copying in...»

«Sons, Seed, and Children of Promise in Galatians: Discerning the Coherence in Paul’s Model of Abrahamic Descent by Bradley R. Trick Department of Religion Duke University Date:_Approved: _ Richard B. Hays, Supervisor _ Douglas Campbell _ Mark Goodacre _ Stephen Chapman Dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Department of Religion in the Graduate School of Duke University 2010 ABSTRACT Sons, Seed, and Children of Promise...»


«    Koliba  CV  Updated:  December  2012       Christopher J. Koliba Work Address: Home Address: 103 Morrill Hall, Burlington, VT 05405 379 Marshall Rd., Duxbury, VT 05676 ckoliba@uvm.edu; 802-656-3772 802-244-4927 EDUCATION 1998. Doctor of Philosophy (Ph.D.). Interdisciplinary Social Science Program. Maxwell School of Citizenship & Public Affairs. Syracuse University. 1992. Master of Public Administration (MPA). Public Administration Department. Maxwell School of Citizenship...»

«CHAUCER AND MORAL PHILOSOPHY: THE VIRTUOUS WOMEN OF THE CANTERBURYTALES By: Denise Baker Baker, Denise. “Chaucer and Moral Philosophy: The Virtuous Women of the Canterbury Tales.” Medium Aevum 60 (1992): 241-256.Made available courtesy of The Society for the Study of Medieval Languages and Literature: http://mediumaevum.modhist.ox.ac.uk/ ***Note: Figures may be missing from this format of the document In The Regement of Princes, Hoccleve insists that Chaucer is not only the equal of Cicero...»

«When Beruriah Met ʿAʾisha: Textual Intersections and Interactions Among Jewish and Muslim Women Engaged with Religious Law by Shari Golberg A thesis submitted in conformity with the requirements for the degree of Doctorate of Philosophy Department for the Study of Religion University of Toronto © Copyright by Shari Golberg 2013 When Beruriah Met ʿAʾisha: Textual Intersections and Interactions Among Jewish and Muslim Women Engaged with Religious Law Shari Golberg Doctorate of Philosophy...»

«PL ATO’S E U T H Y P H RO & M EN O T U TO R I A L R E A D IN G AND E SS AYS Damien Storey 2015 • CO N T E N TS Introduction 2 Reading 2 Essay 2 Some basics of typography 4 Referencing 4 Plagiarism 5 Gobbets 6 Week 1 Socratic philosophy 8 Week 2 Socrates’ search for definitions 9 Week 3 The pious and the god-loved 10 Week 4 All desire is for what’s good I 11 Week 5 All desire is for what’s good II 12 Week 6 Meno’s Paradox & Recollection 13 Week 7 Knowledge and True Belief 14 Week 8...»

«Smart Grid Applicability Prioritisation Of Neighbourhoods by Developing a Geospatial Decision Support Model ZAFER OZTURK College of Science and Technology School of the Built Environment The University of Salford Submitted in Partial Fulfilment of the Requirements of the Degree of Doctor of Philosophy, September 2013 i DEDICATION My angelic mother Gulser Ozturk and my dear father Fahrettin Ozturk taught me how to stand as steady as rock against the facts of life, in the most honourable way....»

«TANGIBLE HOPE: CUBAN PROTESTANTISM IN THE POST-SOVIET ERA By ROSE T. CARAWAY A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2011 1 © 2011 Rose T. Caraway 2 To Mimi and Opa 3 ACKNOWLEDGMENTS This work would not have been possible without the selfless support of countless wonderful people in my life, both inside and outside of academia. Special thanks go to Denise...»

«RELATIONAL TRANSFER IN REINFORCEMENT LEARNING by Lisa Torrey A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy (Computer Sciences) at the UNIVERSITY OF WISCONSIN–MADISON May 2009 i ACKNOWLEDGMENTS This research was partially supported by DARPA grant HR0011-04-1-0007, United States Naval Research Laboratory grant N00173-06-1-G002, and DARPA grant FA8650-06-C-7606. Special thanks to my advisor, Jude Shavlik, to my collaborators Trevor...»

«A Case for Ethical Veganism: Intuitive and methodological considerations. Tristram McPherson Virginia Tech Department of Philosophy tristram@vt.edu Penultimate Manuscript for Journal of Moral Philosophy When the paper is published, it will be available at: http://www.brill.com/journal-moral-philosophy Abstract This paper begins by setting out an intuitive case for ethical veganism: the thesis that it is typically wrong to consume animal products, that begins with the intuitive claim that it is...»

<<  HOME   |    CONTACTS
2016 www.dissertation.xlibx.info - Dissertations, online materials

Materials of this site are available for review, all rights belong to their respective owners.
If you do not agree with the fact that your material is placed on this site, please, email us, we will within 1-2 business days delete him.