Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.
W I S S E N T E C H N I K L E I D E N S C H A F T
contact: thomas.ulz@student.tugraz.at // http://www.ist.tugraz.at/
Peopl...
2
Contents
1. Motivation
2. PeopleViews
3. Recommendation approach
4. Evaluation
5. Ongoing and future work
Alexander Felf...
3
Motivation
Constraint-based Recommendation
Specific type of knowledge based recommendation
Relies on a predefined set of c...
4
Motivation
Knowledge acquisition bottleneck
Only a few Knowledge
Engineers
Possibly a lot of users
with item knowledge
I...
5
Motivation
How willing are end users to contribute?
N=161, 111 would be willing to contribute
[Felfernig et al., CrowdRe...
6
PeopleViews
PeopleViews
Short-term tasks (Micro-tasks)
Domain experts perform short-term knowledge
engineering tasks the...
7
PeopleViews
PeopleViews - Knowledge base
Product attributes
”
Facts“ about items, e.g. sensor size of a camera
Defined wh...
8
PeopleViews
PeopleViews - Features
Users are able to:
Define new knowledge bases
Create new recommendation domain
Add ite...
9
PeopleViews
Definition of knowledge bases
Product attributes
attribute question to user domain
similarity
metric
sensorsi...
10
PeopleViews
Definition of knowledge bases
Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer,...
11
PeopleViews
Definition of knowledge bases
User attributes
attribute choice type question to user domain
usertype multipl...
12
PeopleViews
Definition of knowledge bases
Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer,...
13
PeopleViews
Micro-tasks
choice type: multiple
Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reit...
14
PeopleViews
Micro-tasks
choice type: single
Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiter...
15
PeopleViews
Micro-tasks
choose item, single attribute
Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Ste...
16
PeopleViews
Recommendation view
Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin S...
17
PeopleViews
Recommendation view - Item details
Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Rei...
18
Recommendation approach
Recommendation approach in PeopleViews
User attributes
support(Φ, u, v) =
s(Φ, u, v)
|s(Φ, u, v...
19
Recommendation approach
Recommendation approach in PeopleViews
Product attributes
support(Φ, p, v) =


...
20
Recommendation approach
Recommendation approach in PeopleViews
Selection of recommendation-relevant items
f(Φ) =
u∈U
u ...
21
Evaluation
Evaluation of recommender algorithms
Collect data using WeeVis (http://www.weevis.org)
Canon DSLR recommende...
22
Evaluation
Comparison to other approaches
1 1.5 2 2.5 3 3.5 4 4.5 5
0
10
20
30
40
50
60
70
80
90
100
top n
precisionin%...
23
Ongoing and future work
Ongoing and future work
Recommendation approaches
Implementation and evaluation of further
appr...
24
Thank you!
Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger
Upcoming SlideShare
Loading in …5
×

Peopleviews: Human Computation for Constraint-Based Recommendation

395 views

Published on

CrowdRec2015 Slides

Published in: Data & Analytics
  • Be the first to comment

  • Be the first to like this

Peopleviews: Human Computation for Constraint-Based Recommendation

  1. 1. W I S S E N T E C H N I K L E I D E N S C H A F T contact: thomas.ulz@student.tugraz.at // http://www.ist.tugraz.at/ PeopleViews: Human Computation for Constraint-Based Recommendation Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger, Institute for Software Technology, Graz University of Technology
  2. 2. 2 Contents 1. Motivation 2. PeopleViews 3. Recommendation approach 4. Evaluation 5. Ongoing and future work Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger
  3. 3. 3 Motivation Constraint-based Recommendation Specific type of knowledge based recommendation Relies on a predefined set of constraints Rankings determined by utility function Why constraint-based recommendation? Suitable for complex item domains Possible to ” explain“ recommendations Diagnoses for too strict requirements Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger
  4. 4. 4 Motivation Knowledge acquisition bottleneck Only a few Knowledge Engineers Possibly a lot of users with item knowledge Idea: enable users to contribute to knowledge bases Are users willing to contribute to knowledge bases? End Users Knowledge Engineers Item Knowledge Knowledge Base Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger
  5. 5. 5 Motivation How willing are end users to contribute? N=161, 111 would be willing to contribute [Felfernig et al., CrowdRec 2014] Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger
  6. 6. 6 PeopleViews PeopleViews Short-term tasks (Micro-tasks) Domain experts perform short-term knowledge engineering tasks they are much better in compared to knowledge engineers. Potential advantages Less effort related to recommendation knowledge base development and maintenance Fewer erroneous constraints Significantly higher degree of scalability Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger
  7. 7. 7 PeopleViews PeopleViews - Knowledge base Product attributes ” Facts“ about items, e.g. sensor size of a camera Defined when item is added to knowledge base User attributes Perceived differently by users, e.g. a cameras field of application Defined by users in micro tasks Support Support of item for specific {user, product} attribute value Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger
  8. 8. 8 PeopleViews PeopleViews - Features Users are able to: Define new knowledge bases Create new recommendation domain Add items to existing domains Evaluate existing items ” Answer“ micro tasks Use existing knowledge bases to get recommendations Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger
  9. 9. 9 PeopleViews Definition of knowledge bases Product attributes attribute question to user domain similarity metric sensorsize Preferred sensor size? {fullframe, APS-C, MFT, 1“, 2/3“} EIB max- shutterspeed Required max. shutter speed? {1/4000, 1/6000, 1/8000, 1/16000} LIB maxISO Required max. ISO sensitivity? {6400, 12800, 25600} MIB price Max. price? integer LIB Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger
  10. 10. 10 PeopleViews Definition of knowledge bases Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger
  11. 11. 11 PeopleViews Definition of knowledge bases User attributes attribute choice type question to user domain usertype multiple Suited for whom? { beginner, amateur, expert } application single Preferred { sport, architecture, macro, application? landscape, portrait } usability single Minimum accepted { average, high, very high } usability? Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger
  12. 12. 12 PeopleViews Definition of knowledge bases Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger
  13. 13. 13 PeopleViews Micro-tasks choice type: multiple Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger
  14. 14. 14 PeopleViews Micro-tasks choice type: single Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger
  15. 15. 15 PeopleViews Micro-tasks choose item, single attribute Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger
  16. 16. 16 PeopleViews Recommendation view Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger
  17. 17. 17 PeopleViews Recommendation view - Item details Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger
  18. 18. 18 Recommendation approach Recommendation approach in PeopleViews User attributes support(Φ, u, v) = s(Φ, u, v) |s(Φ, u, v)| · |s(Φ, u, v)| |s(Φ, u)| symbol meaning Φ item u user attribute u ∈ U p product attribute v {user, product} attribute value s(Φ, u, v) support specified by user Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger
  19. 19. 19 Recommendation approach Recommendation approach in PeopleViews Product attributes support(Φ, p, v) =    1if v = val(Φ, p), 0 otherwise EIB 1 − |v - val(Φ, p)| max(Φ, p) - min(Φ, p) NIB val(Φ, p)-min(Φ, p) max(Φ, p) - min(Φ, p) MIB max(Φ, p)-val(Φ, p) max(Φ, p) - min(Φ ,p) LIB Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger
  20. 20. 20 Recommendation approach Recommendation approach in PeopleViews Selection of recommendation-relevant items f(Φ) = u∈U u ∈ values(Φ, u) ∪ {noval} → include(Φ) Ranking items by their utility utility(Φ, REQ) = a=v∈REQ support(Φ, a, v) · w(a) Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger
  21. 21. 21 Evaluation Evaluation of recommender algorithms Collect data using WeeVis (http://www.weevis.org) Canon DSLR recommender 16 items, 7 attributes (27 possible ” answers“) Users defined their requirements and selected best matching camera 356 unique sessions 1 out of N ” training and evaluation“ Is desired item in top n recommended items? Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger
  22. 22. 22 Evaluation Comparison to other approaches 1 1.5 2 2.5 3 3.5 4 4.5 5 0 10 20 30 40 50 60 70 80 90 100 top n precisionin% Learned Weights Support Values Popularity Based Random Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger
  23. 23. 23 Ongoing and future work Ongoing and future work Recommendation approaches Implementation and evaluation of further approaches; diagnoses and repair Micro-task scheduling Automatically assign micro-tasks to users, using a content-based approach Quality assurance Improve dataset quality and prevent manipulation Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger
  24. 24. 24 Thank you! Alexander Felfernig, Thomas Ulz, Sarah Haas, Michael Schwarz, Stefan Reiterer, Martin Stettinger

×