INTRO <ul><li>Social networking is generating an incredible amount of information that is sometimes difficult for users to...
PROPOSAL <ul><li>We propose a Social Mobile Activity Recommender that integrates Facebook social network mobile data and s...
SOcial Mobile Activity Recommender (SOMAR) <ul><li>SOMAR provides recommendations on the basis of the users social relatio...
SOMAR (1) <ul><li>SOMAR can be divided into: i) generation of the recommender and ii) application of the recommender.  </l...
Generation of the recommender <ul><li>The recommendations are based on the user social realtionships, for this reason we p...
Computing the Social Graph <ul><li>The graph computation in the mobile goes on in three steps: </li></ul><ul><li>mutual fr...
Computing the Social Graph (1) <ul><li>Mutual Friend Computation </li></ul><ul><li>Information required: </li></ul><ul><ul...
Computing the Social Graph (2) <ul><li>User Clustering </li></ul><ul><li>The aim is to group the users, the measure of aff...
Computing the Social Graph (3) <ul><li>Affinity Degree Calculation </li></ul><ul><li>In this step we determine the degree ...
Application of the Recommender
ACTIVITY RECOMMENDATION <ul><li>Activity recommendation requires the following online tasks: </li></ul>
ACTIVITY RECOMMENDATION <ul><li>Activity recommendation requires the following online tasks: </li></ul><ul><li>Data Integr...
ACTIVITY RECOMMENDATION <ul><li>Activity recommendation requires the following online tasks: </li></ul><ul><li>Data Integr...
ACTIVITY RECOMMENDATION <ul><li>Activity Matching </li></ul><ul><ul><li>The social graph is used to select activities amon...
CASE STUDY <ul><li>In this paper we only focus in the analysis of the feasibility and the performance of generating the so...
CASE STUDY <ul><li>Data about the target user include the friend list, public wall and all information shared by friends i...
RESULTS (performance analysis) <ul><li>STEP 1 has complexity O(n^m), where n is the number of the Root’s friends and m is ...
RESULTS (performance analysis) (2) <ul><li>STEP 3 has a linear complexity of O(k), where k is the number of identified clu...
CONCLUSIONS <ul><li>SOMAR analyses and compute information locally in the device </li></ul><ul><li>Local computation provi...
CONCLUSIONS <ul><li>The main cost is related to computation of the social graph </li></ul><ul><li>It depends on the number...
Upcoming SlideShare
Loading in...5
×

ISDA 2011 Cordoba

475

Published on

Published in: Technology
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total Views
475
On Slideshare
0
From Embeds
0
Number of Embeds
0
Actions
Shares
0
Downloads
0
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide

ISDA 2011 Cordoba

  1. 1. INTRO <ul><li>Social networking is generating an incredible amount of information that is sometimes difficult for users to process </li></ul><ul><li>Social networking is now supported by mobile devices. The analysis of on-the-move interactions in social networks is now gaining further attention from researchers </li></ul><ul><li>Several links, activities, and recommendations are proposed by networked friends every hour, which together are nearly impossible to manage </li></ul><ul><li>Therefore there is a need to filter and make accessible such information to users </li></ul><ul><li>This is the motivation behind developing a mobile recommender that exploits social network information </li></ul>
  2. 2. PROPOSAL <ul><li>We propose a Social Mobile Activity Recommender that integrates Facebook social network mobile data and sensor data to propose activities (concerts, ongoings, semiars, etc) to the mobile users </li></ul><ul><li>The recommendations are completely calculated in the mobile device with an embedded data mining component </li></ul><ul><li>Endowing mobiles with the capability to generate data mining models it may be possible to create a new smart world! </li></ul>
  3. 3. SOcial Mobile Activity Recommender (SOMAR) <ul><li>SOMAR provides recommendations on the basis of the users social relationships </li></ul><ul><li>Hypothesis: Social networks users tend to have social interactions only with a small group of their social network friends </li></ul><ul><li>Information sources: </li></ul><ul><ul><li>mobile data in terms of call history or contacts </li></ul></ul><ul><ul><li>sensor data in terms of sensor location </li></ul></ul><ul><ul><li>Facebook data </li></ul></ul>
  4. 4. SOMAR (1) <ul><li>SOMAR can be divided into: i) generation of the recommender and ii) application of the recommender. </li></ul>
  5. 5. Generation of the recommender <ul><li>The recommendations are based on the user social realtionships, for this reason we present the Social Graph. </li></ul><ul><li>The aim the Social Graph is: </li></ul><ul><ul><li>to represent the relationships of a user with the users friends or subgroups of friends </li></ul></ul><ul><ul><li>show how frequently the user interacts with them </li></ul></ul><ul><li>The characteristics of the Social graph are: </li></ul><ul><ul><li>the nodes represent friends or groups of friends </li></ul></ul><ul><ul><li>the graph is based on mutual friendship and the quantity of relationships among users </li></ul></ul>
  6. 6. Computing the Social Graph <ul><li>The graph computation in the mobile goes on in three steps: </li></ul><ul><li>mutual friend computation </li></ul><ul><li>user clustering </li></ul><ul><li>affinity degree calculation </li></ul>
  7. 7. Computing the Social Graph (1) <ul><li>Mutual Friend Computation </li></ul><ul><li>Information required: </li></ul><ul><ul><li>the list of friends of the user </li></ul></ul><ul><ul><li>the list of friends of users’s friendes. </li></ul></ul><ul><li>For each pair of friends, the number of mutual friends is calculated </li></ul><ul><li>Output: a matrix M (n*n) of mutual friendship, where n is the numer of the user’s friends. </li></ul>
  8. 8. Computing the Social Graph (2) <ul><li>User Clustering </li></ul><ul><li>The aim is to group the users, the measure of affinity is given by the number of mutual friends </li></ul><ul><li>For this aim we propose to use the hierarchical clustering taking as input the matrix of mutual friendship M </li></ul><ul><li>The output is a number of clusters which represent the connections among friends, independently from the Root user being analyzed </li></ul><ul><li>In fact, a greater number of mutual friends shared by two users results in a higher probability that they will be in the same clusters </li></ul>
  9. 9. Computing the Social Graph (3) <ul><li>Affinity Degree Calculation </li></ul><ul><li>In this step we determine the degree of the affinity between the Root and the clustered friends </li></ul><ul><li>The edges connecting nodes are labeled with a value based on the number of relationships among users, which in turn establishes the degree of affinity between those nodes </li></ul><ul><li>The relationships are defined by: comments, likes, messages etc. </li></ul><ul><li>The final output is a weighed graph representing the social relations of a user. </li></ul>
  10. 10. Application of the Recommender
  11. 11. ACTIVITY RECOMMENDATION <ul><li>Activity recommendation requires the following online tasks: </li></ul>
  12. 12. ACTIVITY RECOMMENDATION <ul><li>Activity recommendation requires the following online tasks: </li></ul><ul><li>Data Integration: </li></ul><ul><ul><li>We will provide a module that integrates and prepares data from sensors, the phone and Facebook </li></ul></ul>
  13. 13. ACTIVITY RECOMMENDATION <ul><li>Activity recommendation requires the following online tasks: </li></ul><ul><li>Data Integration: </li></ul><ul><ul><li>We will provide a module that integrates and prepares data from sensors, the phone and Facebook </li></ul></ul><ul><li>Activity Recognition: </li></ul><ul><ul><li>Activities are easily detectable because they are published by the event Facebook application </li></ul></ul>
  14. 14. ACTIVITY RECOMMENDATION <ul><li>Activity Matching </li></ul><ul><ul><li>The social graph is used to select activities among those detected by the Activity Recognizer </li></ul></ul><ul><ul><li>The weights between the nodes in the social graph represents how socially close is the user proposing the activities to the target (Root node) </li></ul></ul><ul><ul><li>This weight is then compared with a threshold based on the social graphs structure and degree, and it can be modified by the user </li></ul></ul><ul><ul><li>Finally, activities are flaged considering, for example, location and current user activities </li></ul></ul>
  15. 15. CASE STUDY <ul><li>In this paper we only focus in the analysis of the feasibility and the performance of generating the social graph in the mobile, in particular: </li></ul><ul><ul><li>Complexity of step 1 </li></ul></ul><ul><ul><li>Clustering behaviour depending on the number of friends </li></ul></ul><ul><ul><li>Affinity degree calculation of the basis of the number of clusters </li></ul></ul><ul><li>We carried out the experiments in a system with a 2.16 GHz Core 2 processor and 2.5GB 667 MHz DDR2 SDRAM memory </li></ul>
  16. 16. CASE STUDY <ul><li>Data about the target user include the friend list, public wall and all information shared by friends in the friend list </li></ul><ul><li>In particular, the user in this case study has: </li></ul><ul><ul><li>A friend list that contains 149 friends </li></ul></ul><ul><ul><li>Data comprise 1379 posts generated over two years </li></ul></ul><ul><ul><li>Data in post include links, photos, likes,comments and tags that involve two or more users </li></ul></ul>
  17. 17. RESULTS (performance analysis) <ul><li>STEP 1 has complexity O(n^m), where n is the number of the Root’s friends and m is the maximum number of friends connected to each of the Root’s friend. </li></ul><ul><li>STEP 2: CPU time grows up to 13 seconds when the number of friends is maximum </li></ul>
  18. 18. RESULTS (performance analysis) (2) <ul><li>STEP 3 has a linear complexity of O(k), where k is the number of identified clusters. </li></ul>
  19. 19. CONCLUSIONS <ul><li>SOMAR analyses and compute information locally in the device </li></ul><ul><li>Local computation provides several advantages: </li></ul><ul><ul><li>It's guaranteeing the privacy of sensitive information </li></ul></ul><ul><ul><li>It's providing personalized and context aware recommendations </li></ul></ul><ul><li>The main drawback is relates to resource comsuption and performance </li></ul>
  20. 20. CONCLUSIONS <ul><li>The main cost is related to computation of the social graph </li></ul><ul><li>It depends on the number of friends of the user </li></ul><ul><li>Nevertheless, for the user (real) used in the experiments having 149 friends has been shown the feasibility of the method </li></ul><ul><li>For the rest of the steps of the social graph generation, according to the experiments, they are feasible for the maximum number of friends allowed on Facebook </li></ul>

×