relevance of time spent on web pages
play

Relevance of Time Spent on Web Pages WEBKDD August 20, 2006, - PowerPoint PPT Presentation

Mar 02, 2023 •146 likes •276 views

Relevance of Time Spent on Web Pages WEBKDD August 20, 2006, Philadelphia, USA Peter I. Hofgesang hpi@few.vu.nl Intention of an online visitor Real-world: customers have the ability to explicitly express what they are looking for Web:

  1. Relevance of Time Spent on Web Pages WEBKDD August 20, 2006, Philadelphia, USA Peter I. Hofgesang hpi@few.vu.nl

  2. Intention of an online visitor • Real-world: customers have the ability to explicitly express what they are looking for • Web: intention is hidden and can only be partially revealed from implicit indicators in the traces users leave behind WebKDD 2006 Workshop on Knowledge Discovery on the Web, Aug. 20, 2006, at KDD 2006, Philadelphia, PA, USA

  3. (Broadly) Available information • Order of visited pages (P1 � P2 � P3 …) • Page popularity (nr. of times visited) • Time Spent on Page (TSP)? – claimed to be important in IR, HCI, E-learning – only rarely used in WUM – details are often not reported (however, preprocessing is not obvious!) WebKDD 2006 Workshop on Knowledge Discovery on the Web, Aug. 20, 2006, at KDD 2006, Philadelphia, PA, USA

  4. Example I SALES-Woman 26 SALES-Woman 25 SALES-Woman 24 SALES-Woman 23 SALES-Woman 22 SALES-Woman 21 SALES-Woman 20 SALES-Man 19 SALES-Man 18 SALES-Man 17 SALES-Man 16 Pages visited SALES-Man 15 SALES-Man 14 SALES-Man 13 SALES-Man 12 SALES-Rest 11 SHOP-Household 10 SHOP-Household 9 SHOP-Household 8 SHOP-Household 7 SHOP-Household 6 SHOP-Household 5 SALES-Household 4 SALES-Household 3 SALES-Rest 2 Home 1 0 20 40 60 80 100 120 TSoP (seconds) WebKDD 2006 Workshop on Knowledge Discovery on the Web, Aug. 20, 2006, at KDD 2006, Philadelphia, PA, USA

  5. Example II SHOP-PC 11 SHOP-Home 10 SHOP-Child 9 SHOP-Home 8 SHOP-Child 7 Pages visited Info 6 SHOP-PC 5 SHOP-Child 4 SHOP-TV 3 Info 2 SHOP-PC 1 0 20 40 60 80 100 120 140 160 180 200 TSoP (seconds) WebKDD 2006 Workshop on Knowledge Discovery on the Web, Aug. 20, 2006, at KDD 2006, Philadelphia, PA, USA

  6. Example III SHOP-DVD 25 SHOP-DVD 24 SHOP-DVD 23 SHOP-DVD 22 SHOP-DVD 21 SHOP-DVD 20 SHOP-DVD 19 SHOP-DVD 18 SHOP-DVD 17 Home 16 Personal 15 Pages visited Personal 14 Personal 13 Personal 12 Personal 11 Personal 10 Order 9 Order 8 Order 7 CART-Home 6 CART-Add 5 CART-Add 4 CART-Home 3 SHOP-TV 2 Home 1 0 500 1000 1500 TSoP (seconds) WebKDD 2006 Workshop on Knowledge Discovery on the Web, Aug. 20, 2006, at KDD 2006, Philadelphia, PA, USA

  7. Influential factors I TSP 1 = T2 − T1 (optimistic!) • Data preprocessing – filtering out robot transactions – session identification 5 2.2 x 10 Bank 2 Retail 1 • Distraction 1.8 Retail 2 1.6 1.4 Number of clicks 1.2 1 0.8 0.6 0.4 0.2 0 0 10 20 30 40 50 60 70 80 90 100 110 Time (seconds) WebKDD 2006 Workshop on Knowledge Discovery on the Web, Aug. 20, 2006, at KDD 2006, Philadelphia, PA, USA

  8. Influential factors II • Page type (Granularity of pages) • Hierarchy • Network bandwidth and server load • Speed of reading, etc. TSP 2 = T2 − T1 − T networkTraffic − T serverPageGeneration − T distraction WebKDD 2006 Workshop on Knowledge Discovery on the Web, Aug. 20, 2006, at KDD 2006, Philadelphia, PA, USA

  9. Clustering 27.63% 21.18% 19.45% 19.14% 12.6% Catalog 40 40 40 40 40 Contact 39 39 39 39 39 Info 38 38 38 38 38 ThankYou 37 37 37 37 37 Order 36 36 36 36 36 DirectOrder 35 35 35 35 35 CART-Remove 34 34 34 34 34 CART-Change 33 33 33 33 33 CART-Add 32 32 32 32 32 Cadez et al. (2001) CART-Home 31 31 31 31 31 Search 30 30 30 30 30 SALES-Rest 29 29 29 29 29 SALES-Games 28 28 28 28 28 SALES-TV 27 27 27 27 27 SALES-DVD 26 26 26 26 26 SALES-GSM 25 25 25 25 25 SALES-PC 24 24 24 24 24 SALES-Sport 23 23 23 23 23 PageId SALES-Household 22 22 22 22 22 SALES-Home 21 21 21 21 21 SALES-Beauty 20 20 20 20 20 SALES-Child 19 19 19 19 19 SALES-Man 18 18 18 18 18 SALES-Woman 17 17 17 17 17 SHOP-Rest 16 16 16 16 16 SHOP-Special 15 15 15 15 15 SHOP-Games 14 14 14 14 14 SHOP-TV 13 13 13 13 13 SHOP-DVD 12 12 12 12 12 SHOP-GSM 11 11 11 11 11 SHOP-PC 10 10 10 10 10 SHOP-Sport 9 9 9 9 9 SHOP-Household 8 8 8 8 8 SHOP-Home 7 7 7 7 7 SHOP-Beauty 6 6 6 6 6 SHOP-Child 5 5 5 5 5 SHOP-Man 4 4 4 4 4 SHOP-Woman 3 3 3 3 3 Personal 2 2 2 2 2 Home 1 1 1 1 1 0 0.1 0.2 0 0.5 0 0.1 0.2 0 0.1 0.2 0 0.2 0.4 33.03% 22.01% 19.56% 17.63% 7.78% Catalog 40 40 40 40 40 Contact 39 39 39 39 39 Info 38 38 38 38 38 ThankYou 37 37 37 37 37 Order 36 36 36 36 36 DirectOrder 35 35 35 35 35 CART-Remove 34 34 34 34 34 CART-Change 33 33 33 33 33 CART-Add 32 32 32 32 32 CART-Home 31 31 31 31 31 Search 30 30 30 30 30 SALES-Rest 29 29 29 29 29 SALES-Games 28 28 28 28 28 SALES-TV 27 27 27 27 27 SALES-DVD 26 26 26 26 26 SALES-GSM 25 25 25 25 25 SALES-PC 24 24 24 24 24 SALES-Sport 23 23 23 23 23 PageId SALES-Household 22 22 22 22 22 SALES-Home 21 21 21 21 21 SALES-Beauty 20 20 20 20 20 SALES-Child 19 19 19 19 19 SALES-Man 18 18 18 18 18 SALES-Woman 17 17 17 17 17 SHOP-Rest 16 16 16 16 16 SHOP-Special 15 15 15 15 15 SHOP-Games 14 14 14 14 14 SHOP-TV 13 13 13 13 13 SHOP-DVD 12 12 12 12 12 Similarity based SHOP-GSM 11 11 11 11 11 SHOP-PC 10 10 10 10 10 SHOP-Sport 9 9 9 9 9 SHOP-Household 8 8 8 8 8 SHOP-Home 7 7 7 7 7 SHOP-Beauty 6 6 6 6 6 SHOP-Child 5 5 5 5 5 SHOP-Man 4 4 4 4 4 SHOP-Woman 3 3 3 3 3 Personal 2 2 2 2 2 Home 1 1 1 1 1 0 0.1 0.2 0 0.5 1 0 0.5 0 0.2 0.4 0 0.5 1

  10. Conclusion • TSP is a sensitive measure • Web log data preprocessing and Time normalization required • Added value in identifying user intention • For many applications the combination of TSP and frequency may be the optimal choice WebKDD 2006 Workshop on Knowledge Discovery on the Web, Aug. 20, 2006, at KDD 2006, Philadelphia, PA, USA

  11. Future (current) work • Objective measures of relevance • Normally field experiment to provide some kind of labeled data • Special testbed � – e.g., in case of a retail shop environment we have special labels for buyers – the purchased items indicate user interest and can be compared with the visit WebKDD 2006 Workshop on Knowledge Discovery on the Web, Aug. 20, 2006, at KDD 2006, Philadelphia, PA, USA

  12. Questions?

Download Presentation
Download Policy: The content available on the website is offered to you 'AS IS' for your personal information and use only. It cannot be commercialized, licensed, or distributed on other websites without prior consent from the author. To download a presentation, simply click this link. If you encounter any difficulties during the download process, it's possible that the publisher has removed the file from their server.

Recommend

CSE 373: P vs NP Michael Lee Monday, Mar 5, 2018 1 Overview Previously: We spent a lot of

CSE 373: P vs NP Michael Lee Monday, Mar 5, 2018 1 Overview Previously: We spent a lot of

CSE 373: P vs NP Michael Lee Monday, Mar 5, 2018 1 Overview Previously: We spent a lot of time analyzing algorithms 2 We spent a lot of time learning how to solve problems Overview Previously: 2 We spent a lot of time learning how

1.23k views • 122 slides
Topic of this talk Topic of this talk From E- -Relevance Relevance From E to W- -Relevance

Topic of this talk Topic of this talk From E- -Relevance Relevance From E to W- -Relevance

Topic of this talk Topic of this talk From E- -Relevance Relevance From E to W- -Relevance Relevance to W What is relevance in Mobile Personal Information Retrieval? P. Coppola, V. Della Mea, L. Di Gaspero, S. Mizzaro What is

65 views • 6 slides
Time Spent With Media Time Spent With Media Average amount of time spent with each medium

Time Spent With Media Time Spent With Media Average amount of time spent with each medium

Time Spent With Media Time Spent With Media Average amount of time spent with each medium Average amount of time spent with each medium per day, among 8- -18 year 18 year- -olds: olds: per day, among 8 3:51 3:51 Total Total 4

852 views • 29 slides
SYMBOLS FOR TIME Cost spent to build variation point i at time = time variable t = time now,

SYMBOLS FOR TIME Cost spent to build variation point i at time = time variable t = time now,

SYMBOLS FOR TIME Cost spent to build variation point i at time = time variable t = time now, T T = target date target date T* = modeling limit (t=forever) i = index over variation points SYMBOLS FOR TIME Cost spent to build variation point i

285 views • 16 slides
sIT ossIG J Last N 500 pages Ps 80 time S Relation 100 M 1000 pages Pr R Nested

sIT ossIG J Last N 500 pages Ps 80 time S Relation 100 M 1000 pages Pr R Nested

sIT ossIG J Last N 500 pages Ps 80 time S Relation 100 M 1000 pages Pr R Nested LoopJoions w R CPR M Nam Simple outer Page oriented M N 1M buffer K M1 Block based w m Index data access costumed index access cost MTM.pp IB't

509 views • 6 slides
TANNING: Since you spent a ton of money, time and energy preparing for your competition, my

TANNING: Since you spent a ton of money, time and energy preparing for your competition, my

POSING AND PRESENTATION 101 TANNING: Since you spent a ton of money, time and energy preparing for your competition, my suggestion is to spend the extra couple dollars and get professionally tanned. There are services now that will get you

539 views • 6 slides
Parents evening presentation We have spent a great deal of time recently talking to the

Parents evening presentation We have spent a great deal of time recently talking to the

Parents evening presentation We have spent a great deal of time recently talking to the students about their futures UCAS and University. Essex. Kent. LSE. QMUL. Cardiff. Studying overseas. Higher Degree Apprentiships. Work. School

402 views • 26 slides
OPTIMISE THE VALUE OF EVERY POUND SPENT ON YOUR UTILITIES Brought to you by CRYSTAL BALL TIME

OPTIMISE THE VALUE OF EVERY POUND SPENT ON YOUR UTILITIES Brought to you by CRYSTAL BALL TIME

OPTIMISE THE VALUE OF EVERY POUND SPENT ON YOUR UTILITIES Brought to you by CRYSTAL BALL TIME The clearest picture of the energy market WIDER VISIBILITY OF WHOLESALE ENERGY PRICES Newspapers Generators Suppliers Other traders EXTERNAL

141 views • 10 slides
PG04 | Personal Tutorial For my PG04 Concept and Prototyping unit, Ive spent some time

PG04 | Personal Tutorial For my PG04 Concept and Prototyping unit, Ive spent some time

Thursday 16th May PG04 | Personal Tutorial For my PG04 Concept and Prototyping unit, Ive spent some time exploring some aspects of creative programming and interacting with real objects, such as lights, lasers and machines. It's not

492 views • 15 slides
A Foundational Leadership Strategy Some Data About Listening 80% of awake time is spent

A Foundational Leadership Strategy Some Data About Listening 80% of awake time is spent

Listening A Foundational Leadership Strategy Some Data About Listening 80% of awake time is spent communicating We spend 45% of awake time listening 75 _____% of the words are ignored, misunderstood, or forgotten Most

558 views • 20 slides
Loops Most execution time in most programs is spent in loops: 90/10 is typical CS 4120

Loops Most execution time in most programs is spent in loops: 90/10 is typical CS 4120

Loops Most execution time in most programs is spent in loops: 90/10 is typical CS 4120 Most important targets of optimization: loops Introduction to Compilers Loop optimizations: loop-invariant code motion Andrew Myers

459 views • 3 slides
Time Spent on Research with Undergraduate Students Gender Differences among Gender Differences

Time Spent on Research with Undergraduate Students Gender Differences among Gender Differences

Time Spent on Research with Undergraduate Students Gender Differences among Gender Differences among STEM Faculty Amber D. Lambert Amy K. Garver Allison BrckaLorenz Indiana University Antwione Haywood Center for Postsecondary Research

235 views • 12 slides
Let's talk about time spent working ? May 2014 CONTENTS 1. Method and context 2. Stability of

Let's talk about time spent working ? May 2014 CONTENTS 1. Method and context 2. Stability of

Let's talk about time spent working ? May 2014 CONTENTS 1. Method and context 2. Stability of expectations and challenges related to the management of professional development 3. The organisation of working hours: the new "frontier" to be

346 views • 22 slides
Monday, 06 March 2017 To the attention of Dear, While thanking you in advance for the time spent

Monday, 06 March 2017 To the attention of Dear, While thanking you in advance for the time spent

Monday, 06 March 2017 To the attention of Dear, While thanking you in advance for the time spent in reading this letter, we are proud to introduce the winery Le Tenute di Carlo Pili, located in Soleminis, Cagliari, Italy. Le Tenute Carlo

452 views • 10 slides
1 Interim Spent Fuel Storage Facility (ISF-2) ISF-2 Characteristics Storage of 21,297 Spent

1 Interim Spent Fuel Storage Facility (ISF-2) ISF-2 Characteristics Storage of 21,297 Spent

1 Interim Spent Fuel Storage Facility (ISF-2) ISF-2 Characteristics Storage of 21,297 Spent Fuel assemblies (SFA) 2,500 assemblies processed and stored per year Spent Fuel Storage Area (SFSA) includes 232 Concrete Storage Modules

359 views • 21 slides
Honor Your Mentor Honor Your Mentor W. Allen Addison, MD , The time you spent with us and the

Honor Your Mentor Honor Your Mentor W. Allen Addison, MD , The time you spent with us and the

Honor Your Mentor Honor Your Mentor W. Allen Addison, MD , The time you spent with us and the teaching you bestowed upon us teaching you bestowed upon us changed our lives. We love you and are forever deeply indebted. Robert L. Harris, MD;

517 views • 30 slides
The Lifecyle of a Youtube Video: Phases, Content and Popularity Honglin Yu, Lexing Xie, Scott

The Lifecyle of a Youtube Video: Phases, Content and Popularity Honglin Yu, Lexing Xie, Scott

The Lifecyle of a Youtube Video: Phases, Content and Popularity Honglin Yu, Lexing Xie, Scott Sanner Australian National University, NICTA May 22, 2015 Overview The scarce, and therefore valuable, resource is now attention B. A.

410 views • 13 slides
Clustering in Popularity Adjusted Stochastic Block Model Majid Noroozi and Marianna Pensky

Clustering in Popularity Adjusted Stochastic Block Model Majid Noroozi and Marianna Pensky

Clustering in Popularity Adjusted Stochastic Block Model Majid Noroozi and Marianna Pensky Department of Mathematics University of Central Florida Majid Noroozi (UCF) Clustering in PABM 1 / 25 Introduction Clustering is a central problem in

503 views • 37 slides
COPING WITH THE CHALLENGE OF SORTING LARGE PRODUCT CATALOGS ONLINE - SHOP WINDOW ARRANGEMENT

COPING WITH THE CHALLENGE OF SORTING LARGE PRODUCT CATALOGS ONLINE - SHOP WINDOW ARRANGEMENT

Please write title, subtitle and speaker name in all capital letters COPING WITH THE CHALLENGE OF SORTING LARGE PRODUCT CATALOGS ONLINE - SHOP WINDOW ARRANGEMENT CAGDAS SENOL 19-06-2019 Put images in the grey Put images in the grey

760 views • 45 slides
Data Streams Many large sources of data are generated as streams of updates: IP Network

Data Streams Many large sources of data are generated as streams of updates: IP Network

Summarizing and Mining Skewed Data Streams Graham Cormode cormode@bell-labs.com Flip Korn, S. Muthukrishnan, Divesh Srivastava Data Streams Many large sources of data are generated as streams of updates: IP Network traffic data Text:

485 views • 44 slides
The Dynamics of Repeat Consumption Ashton Anderson Stanford University Ravi Kumar, Andrew

The Dynamics of Repeat Consumption Ashton Anderson Stanford University Ravi Kumar, Andrew

The Dynamics of Repeat Consumption Ashton Anderson Stanford University Ravi Kumar, Andrew Tomkins, Sergei Vassilvitskii Google Thursday, April 10, 14 repeat consumption a lot of consumption is repeat consumption what factors determine what

835 views • 64 slides
Browser Feature Usage on the Modern Web Summary Analysis of how frequently javascript

Browser Feature Usage on the Modern Web Summary Analysis of how frequently javascript

Browser Feature Usage on the Modern Web Summary Analysis of how frequently javascript features are used Feature defined as browser capability accessed through JavaScript function or property. Considers only sites in Alexa

523 views • 11 slides
Shallow Reading with Deep Learning Predicting popularity of online content using only its title W.

Shallow Reading with Deep Learning Predicting popularity of online content using only its title W.

Shallow Reading with Deep Learning Predicting popularity of online content using only its title W. Stokowiec, W Trzci ski, K. Wo k, K. Marasek and P. Rokita Polish-Japanese Academy of Information Technology, Warsaw University of

491 views • 20 slides
CSE 154 LECTURE 4: FLOATING AND POSITIONING The CSS float property property description float

CSE 154 LECTURE 4: FLOATING AND POSITIONING The CSS float property property description float

CSE 154 LECTURE 4: FLOATING AND POSITIONING The CSS float property property description float side to hover on; can be left, right, or none (default) a floating element is removed from normal document flow underlying text wraps around

495 views • 15 slides

More recommend