1 contact prediction, routing and fast information spreading in social networks kazem jahanbakhsh...
TRANSCRIPT
1
Contact Prediction, Contact Prediction, Routing and Fast Routing and Fast
Information Information Spreading in Social Spreading in Social
NetworksNetworksKazem Jahanbakhsh
Computer Science DepartmentUniversity of Victoria
August 2012
2
OutlineOutline• Problem Definition and the Context• Routing in Mobile Social Settings• Human Mobility and Contact Event• Collecting Contact Data• Contact Prediction• Hidden Contact Prediction• Fast Information Spreading• Conclusions, Major Contributions and Future Work
3
Problem DefinitionProblem Definition
Message routing, human contact prediction and fast information spreading in the context of human social networks.
4
Routing in Mobile Social Settings
• Motivation: First empirical evaluation of Milgram's experiment in mobile settings
• Social Profile: Set of social characteristics for a user:o Affiliation, Hometown, Language, Nationality, Interests and so on
• Goal: Designing an efficient routing algorithm• Efficiency: Minimizing message forwardings &
Maximizing the probability of message delivery• Assumptions & Constraints:
• Message delivery in physical proximity• Sender knows the destination social profile
5
Social-Greedy Routing Algorithms
• Approach: a greedy strategy by computing similarities between people social profileso Social-Greedy I: Sender forwards the message “m” to nodes socially
closer to destination.o Social-Greedy II & III: Variations of Social-Greedy I.
• Our work is different from previous work because we only make use of social profiles of people for routing!
• Real Data: Infocom 2006 contact trace - 79 people - a brief version of social profiles
6
SDR & CostSDR & Cost
Performance Results for Different Routing Schemes (TTL=9h)
7
Human Mobility & Human Mobility & Contact DataContact Data
Kenny
Eric
Eric Kenny 10:00AM 10:10AM
Kenny Eric 10:00AM 10:10AM
Contact Event: 10:00-10:10 AM
7
8
Contact GraphsContact GraphsEric Butters
Kenny Sara
Katy
Jack
Kyle
9
Collecting Data from Different Collecting Data from Different
Social SettingsSocial Settings
10
Real Data DescriptionsReal Data DescriptionsDataset Inf 05 Inf 06 MIT Camb Roller
Sensors 41 79 97 36 62
Length 3 days 4 days 246 days 11 days 3 hours
Scanning Time
120 sec 120 sec 300 sec 600 sec 15 sec
Ext. Nodes 206 4321 20698 11367 1050
Total Cont. 227657 28216 285512 41587 132511
Ext. Cont. 57056 5757 183135 30714 72365
Ext. Cont. % 25% 20% 64% 74% 55%
Dataset No. of Nodes No. of Edges
Facebook 63731 817090
11
Contact Prediction: Problem
Definition and Assumptions
12
Social Information & Small-World Network Properties
• Birds of a Feather (Homophily)• Using Social Profiles:
o Jacard Social Similarity (Jac)o Social Foci Similarity (Foci)o Max Social Similarity (Max)
• Using Contact Graphs:o Transitivity:
• Number of Common Neighbors (NCN)o Low Diameter :
• Shortest Path (SP) • Random Walk (RW)
• How to reconstruct?
13
Contact Prediction Contact Prediction ResultsResults
Infocom 2006
14
Hidden Contact Hidden Contact PredictionPrediction
15
Hidden Contact Hidden Contact Prediction: Prediction:
Reconstruction Reconstruction AlgorithmAlgorithm
• Methods:o Time-Spatial Locality: NCN, Jacard & MINo Contact Rates: Popularityo Social Similarity: Foci & Jacardo Social Similarity-NCN: Foci-NCN
• Algorithm:• For each compute and store quadruples
in• Sort in a descending order using similarity
scores• Output the first number of quadruples
16
Hidden Contact Hidden Contact Prediction ResultsPrediction Results
Infocom 2006
17
Supervised Learning Supervised Learning ApproachApproach
• Techniques:o Logistic Regressiono K-Nearest Neighbor (KNN)
• Extracted Features: o Contact Graph-based (Degree, Product of degrees, NCN) o Contact Durationo Social Profileso Static Sensors
Session Type Keynote Lunch Break Coffee Break
TPR 0.18/0.24 0.37/0.40 0.41/0.43
FPR 0.03/0.08 0.04/0.07 0.02/0.02
Accuracy 81%/78% 84%/81% 92%/92%
RMSE 0.42/0.40 0.39/0.36 0.26/0.24
Prediction Results (Logistic Regression/KNN)
17
18
• Input: social graph G=(V,E) & a unique message for each node
• Communication Model: synchronized• Constraints: no global information & one contact
per round• Termination: when every node receives all
messages• Goal: analyzing running times of three
information spreading algorithms
Fast Information Spreading Fast Information Spreading in Social Networksin Social Networks
18
19
Information Spreading Information Spreading AlgorithmsAlgorithms
• Random push-pull: o In each round, every node randomly chooses one of its neighbors for
message exchange
• Doerr: o In each round, every node randomly chooses one of its neighbors
except the one that has been just contacted
• Censor: Hybrid strategy:o Even rounds: each node runs random push-pullo Odd rounds: each node chooses one of its neighbors in a sequential
manner from its Bottleneck List
20
Empirical Results from Facebook Graph
Running Times Without 1-whiskersRunning Times on Original Facebook Graph 20
21
Conclusions & Future Conclusions & Future WorkWork
• Major Contributions:
• Social-Greedy Algorithm: o Suitable for bootstrapping wireless devices
• Contact Prediction:o Social Similarity methods, SP and RW outperform randomo Foci-NCN provides the best precision resultso Supervised learning is an effective technique for contact prediction
• Information spreading: o Censor performs well for spreading information in social networks
• Future Work:o Proposing more efficient predictors for large geographical spaceso Final Goal: Predicting where people go next and who they will meet there!
22
Hidden Contacts Hidden Contacts Prediction ResultsPrediction Results
MIT Campus 22
3 4 5 6 7 8 9 10 11 120
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
0.5Performance Evaluation (no of external nodes = 73)
log2 Rank
The
Per
cent
age
of T
rue
Posi
tives
NCN
Jac
Min
Pop
Rand
23
Supervised Learning ResultsSupervised Learning Results
Session Type
Keynote Lunch Break
Coffee Break
degree 4 5 5
degree 7 7 7
degree prod.
3 3 6
ncn 1 1 2
total overlap
2 2 1
social 5 6 4
ncsn 6 4 3Ranking Features
23
24
Examples of 1-Examples of 1-WhiskersWhiskers
24