CS 2650 DISTRIBUTED MULTIMEDIA SYSTEMS SOCIAL NETWORK ANALYSIS
CS 2650 DISTRIBUTED MULTIMEDIA SYSTEMS SOCIAL NETWORK ANALYSIS -LINK RELATIONSHIPS Wei-Hsuan Lin December 2, 2010
OVERVIEW 2010/12/02 Definition Measures Researches Supportive Interactions BBS � Nordic CS 2650 � Online academic web Web space Project 2
DEFINITION 2010/12/02 Social network analysis � Social relationships in terms of network theory CS 2650 � Nodes individual actors within the networks � Links ties, edges, connections the relationships between the actors 3
MEASURES 2010/12/02 Size Density/Connectedness Clique Centralization/Degree, Betweenness k-Cores Distance Diameter CS 2650 4
RESEARCHES (CHANG, 2009) ONLINE SUPPORTIVE INTERACTIONS 2010/12/02 Bulletin Board System (BBS) � users post messages for others to read and reply to CS 2650 � PTT the most popular in Taiwan the largest BBS in the Chinese-speaking community 5
METHOD (CHANG, 2009) 2010/12/02 Post classification � Threads CS 2650 At least on initiator and one reply � Isolates receive no replies Tool � UCINET 6
MEASURES (CHANG, 2009) Size � Density/connectedness � Number of individuals involved in exchange linkages in a network CS 2650 2010/12/02 Ratio of the number of actual communication links to the number of possible communication links in the network Clique ≥ 3 individuals � All individuals in a clique are connected � no individuals have communication links to individuals outside the clique � Centralization � Degree: the number of direct links with others 7
RESULT (CHANG, 2009) 2010/12/02 CS 2650 8
RESULT (CHANG, 2009) 2010/12/02 Sparse communication networks � Low density CS 2650 9
RESULT (CHANG, 2009) 2010/12/02 CS 2650 high centralization Some people were always providing support some were always on the receiving side 10
RESULT (CHANG, 2009) small cliques 2010/12/02 CS 2650 11
RESULT (CHANG, 2009) 2010/12/02 CS 2650 Participation over time was rather transitory Individuals just came in, wrote posts, waited for replies, and left. 12
Research (Ortega & Aguillo, 2008) Nordic academic web Visualize link relationships in the Nordic academic web space Finnish, 11 Danish, 28 Swedish academic web domain � Select from Webometrics Ranking of World Universities (www. webometrics. org) � http: //www. webometrics. info/index. html CS 2650 � 23 2010/12/02 13
METHOD(ORTEGA & AGUILLO, 2008) Graph through link matrix � the Tool � Net. Draw 2. 2 CS 2650 main characteristics of the network 2010/12/02 Integrated with UCINET for drawing diagrams of social networks 14
MEASURE(ORTEGA, AGUILLO, 2008) Degree � Betweenness � p-Cliques � � sub-network in which each node has k degree in that sub-network Detect group with a strong link density Distance � � sub-network where every node is directly connected among them Group with a high density and a way to detect underlying sub-networks k-Cores � � Hubs or gateways that connect different web networks CS 2650 Incoming links and outcoming links 2010/12/02 the number of steps in the shortest path that connect two nodes The distance is short, there is a strong cohesion Diameter � the number of steps in the longest path 15
RESULT(ORTEGA, AGUILLO, 2008) Strongly connected network � 2. 16 average distance between nodes 2010/12/02 CS 2650 16
RESULT(ORTEGA, AGUILLO, 2008) Almost the same universities 2010/12/02 CS 2650 17
RESULT(ORTEGA, AGUILLO, 2008) � Mediator between sub-networks � Centre of the Nordic university web space CS 2650 The cohesion is supported by some gateway universities or hubs that mediate between subnetworks 2010/12/02 18
MY PROJECT PTT � Shuang. He Measure � size, density, clique, centralization-degree CS 2650 2010/12/02 UCINET, Net. Draw collect the data manually Three months � Dynamics by observation of each month PIPE 2 19
MODELLING 2010/12/02 CS 2650 20
REFERENCE CS 2650 Chang, H. -J. (2009), Online supportive interactions: Using a network approach to examine communication patterns within a psychosis social support group in Taiwan. Journal of the American Society for Information Science and Technology, 60(7): 1504– 1517. doi: 10. 1002/asi. 21070 Ortega, J. L. and Aguillo, I. F. (2008), Visualization of the Nordic academic web: Link analysis using social network tools. Information Processing and Management, 44(4): 1624 -1633. doi: 10. 1016/j. ipm. 2007. 09. 010 UCINET http: //www. analytictech. com/ucinet/ Social Network http: //en. wikipedia. org/wiki/Social_network 2010/12/02 21
- Slides: 21