Machine Learning and Data Mining Methods for Recommender ...
A recommender system for social learning platforms
-
Upload
soude-fazeli -
Category
Technology
-
view
357 -
download
2
description
Transcript of A recommender system for social learning platforms
![Page 1: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/1.jpg)
page 1
![Page 2: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/2.jpg)
page 2
![Page 3: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/3.jpg)
page 3
Recommender systems?
![Page 4: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/4.jpg)
page 4
4
A socially-powered, multilingual open learning infrastructure in Europe
Open Discovery Space (ODS)
Recommendations!
Which algorithm best fits ODS platform?
![Page 5: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/5.jpg)
page 5
Towards a Trust-based Recommender for Social platforms
Soude Fazeli, PhD candidate Dr. Hendrik Drachsler Prof. Dr. Peter Sloep
![Page 6: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/6.jpg)
page 6
6
1. Content-based 2. Collaborative filtering ✓
Recommender algorithms
![Page 7: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/7.jpg)
page 7
7
Collaborative filtering algorithms
![Page 8: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/8.jpg)
page 8
Sparsity!
Similarity
![Page 9: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/9.jpg)
page 9
Improving prediction accuracy of recommendations
Trustworthy users == like-minded users !
![Page 10: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/10.jpg)
page 10
• Golbeck’s TidalTrust • Trust-aware recommender by Massa and Avesani • Andersen et al’s axiomatic approach • T-BAR by Bellaachia and Alathel • And many more…
All require users to give explicit trust ratings!
![Page 11: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/11.jpg)
page 11
Centralized approaches à scalability problem
• Lathia et al.’s trust-based recommender #neal-lathia #recsys
• Trust model by O’Donovan and Smyth
![Page 12: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/12.jpg)
page 12
A social recommender system: T-index approach 1. Description
• Trust networks: a graph • Nodes: users • Edges: trust relationships • Weights: trust values originating from similarity
• Each user can be assumed as an agent • Improve the process of finding nearest neighbors
• T-index • TopTrustee
![Page 13: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/13.jpg)
page 13
• A new trust relationship between two far unconnected users is
inferred if and only if: • Condition 1:
• Mutual trust value between intermediate users is higher than a certain threshold (v)
• Condition 2: • The number of connecting edges is lower than an upper bound
(L)
A social recommender system: T-index approach 2. Trust propagation mechanism
![Page 14: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/14.jpg)
page 14
Alice
Carol
Bob
rated rated
rated
rated
if A trusts B and B trusts C, then A trusts C if and only if condition 1 is met
and condition 2 is met
![Page 15: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/15.jpg)
page 15
• T-index: measure of users’ trustworthiness • H-index: the impact of publications of an author
Indegree (ua) = 7 Indegree (ub) = 5 T-index (ua ) = 2 T-index (ub ) = 4
A social recommender system: T-index approach 3. T-index? Note! Cluster: a group of users
who all trust a common user as the most trustworthy one (central user)
![Page 16: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/16.jpg)
page 16
A social recommender system: T-index approach 3. T-index?
![Page 17: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/17.jpg)
page 17
A social recommender system: T-index approach 4. What T-index is for? • TopTrustee : a list of top raters of an item sorted by T-index • Helps the process of finding nearest neighbors
• Providing access to trustworthy users across the trust network including even those outside the traversal path length limit (L)
![Page 18: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/18.jpg)
page 18
Social data
![Page 19: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/19.jpg)
page 19
• RQ1: How to generate more accurate and thus,
more relevant recommendations by using the social data originating from social activities of users within an online environment?
• RQ2: Can the use of the inter-user trust
relationships that originate from the social activities of users within an online environment further evolve the network of users?
![Page 20: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/20.jpg)
page 20
Hypothesis
![Page 21: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/21.jpg)
page 21
• H2: Using T-index-based trust relationships
between users improves the structure of the trust network by providing balanced degree distribution of the users.
• H1: The extended T-index algorithm outperforms
the classical collaborative filtering algorithms in terms of F1 score, which is a combined value of precision and recall of the generated recommendations.
![Page 22: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/22.jpg)
page 22
Data-driven study 1. Goal
To find out which recommender algorithms best performs and thus, is suitable for social online platforms like ODS platform
![Page 23: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/23.jpg)
page 23
Data-driven study 2. Method
• Testing several recommender algorithms • Classical collaborative filtering algorithms using traditional
nearest neighbors method • T-index approach using graph-based method
• Datasets • MovieLens – standard dataset • MACE, OpenScout, Travel well -- similar to the future ODS
dataset • Using Mahout
![Page 24: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/24.jpg)
page 24
Data-driven study 3. Setting • v = 0.1 (Condition 1), L = 2 (Condition 2) • Training set 80% and test set 20% • Sizes of neighborhoods n= (3,5,7,10) • Size of TopTrustee list m=5
![Page 25: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/25.jpg)
page 25
Data-driven study 4. Result 4.1. F1 score
F1 of the extended T-index and Tanimoto algorithms for different datasets, based on the size of neighborhood
0"0.01"0.02"0.03"0.04"0.05"0.06"0.07"0.08"0.09"0.1"
3" 5" 7" 10"
F1@10%
size%of%neighborhood%(n)%
MACE%
Tanimoto4Jaccard"(CF1)"
Loglikelihood"(CF2)"
Euclidean"(CF3)"
Graph4based"(CF4)"
0"
0.02"
0.04"
0.06"
0.08"
0.1"
0.12"
0.14"
3" 5" 7" 10"
F1@10%
size%of%neighborhood%(n)%
OpenScout%
Tanimoto3Jaccard"(CF1)"
Loglikelihood"(CF2)"
Euclidean"(CF3)"
Graph3based"(CF4)"
0"
0.02"
0.04"
0.06"
0.08"
0.1"
3" 5" 7" 10"
F1@10%
size%of%neighborhood%(n)%
Travel%well%
Tanimoto3Jaccard"(CF1)"
Loglikelihood"(CF2)"
Euclidean"(CF3)"
Graph3based"(CF4)"
0"
0.05"
0.1"
0.15"
0.2"
0.25"
3" 5" 7" 10"
F1@10%
size%of%neighborhood%(n)%
MovieLens%
Tanimoto0Jaccard"(CF1)"
Loglikelihood"(CF2)"
Euclidean"(CF3)"
Graph0based"(CF4)"
![Page 26: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/26.jpg)
page 26
Data-driven study 4.2. Created trust network
Without T-index With T-index
![Page 27: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/27.jpg)
page 27
2. Data-driven study 4.3. Degree centrality
Degree distribution of top-10 central users for different datasets
0"
50"
100"
150"
200"
250"
u1" u2" u3" u4" u5" u6" u7" u8" u9" u10"
degree%
Top)10%central%users%
MovieLens"
OpenScout"
MACE"
Travel"well"
![Page 28: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/28.jpg)
page 28
Conclusion
• The aim is to support user in social platforms to find the most suitable content or people
• Recommender systems can be a solution • Using trust-based approaches to improve
performance of recommender systems even in case of sparsity
![Page 29: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/29.jpg)
page 29
Ongoing and Further work
• Testing recommender algorithms on more datasets coming from social networking sites
• Go online with the ODS platform (October 2013) • User evaluation study (February 2013)
![Page 30: A recommender system for social learning platforms](https://reader035.fdocuments.net/reader035/viewer/2022081403/554f0cb4b4c90547648b552b/html5/thumbnails/30.jpg)
page 30
Soude Fazeli PhD candidate Open University of the Netherlands Centre for Learning Sciences and Technologies (CELSTEC) PO-‐Box 2960 6401 DL Heerlen, The Netherlands email: [email protected]