Schwartz, M., Wood, D.: Discovering shared interests using graph analysis. Communications of the ACM 36 (1993) 78--89  Home/Search   Document Not in Database   Summary   Related Articles   Check

....of social networks. There have been more recent efforts directed at leveraging social networks algorithmically for diverse purposes such as expertise location [17] detecting fraud in cellular communications [9] and mining the network value of customers [8] In particular, Schwartz and Wood [26] construct a graph using email as links, and analyze the graph to discover shared interests. While their domain (like ours) consists of interactions between people, their links are indicators of common interest, not antagonism. Related research includes work on incorporating the relationship ....

....postings were either not responses, or we were not able to find or match the name of the original poster. Because of the sampled nature of the data, if we were to form the graph corresponding to all the responses, parts of the graph would contain too little information to be analyzed effectively [26]. Each data set contained a core connected component (see Figure 3) that comprised almost all the postings within the data set. We, therefore, omitted all vertices (and corresponding documents) that do not connect to the core component. Figure 3 also shows the total number of authors and the ....

M. F. Schwartz and D. C. M. Wood. Discovering shared interests using graph analysis. Communications of the ACM, 36(8):78--89, 1993.

....[3] The success of the social network in part is due to the six degrees of separation or small world phenomenon where any two individuals are separated by a small number of direct personal relationships. For example, within the electronic community, analysis of email logs by Schwartz and Wood [12] indicate that the average distance between any two persons is 5:4. The limitation on publicly available information also contributes to the ecacy of personal networks. A person cannot record entirely the knowledge of his expertise, and often is reluctant to answer queries from strangers. A ....

Schwartz, M. F. and Wood, D. C. M. (1993) \Discovering shared interests using graph analysis," Communications of ACM, vol: 36, no: 8, pp. 78-89. 47

....have been reported so far. The earliest notable systems are HelpNet [11] and Expert Expert Locator(EEL) 18] 19] 11] uses the user filled information to provide the expertise profile and [18] 19] use a representative collection of user s technical document to build the expertise index. [15]uses graph analysis to find shared interest by analyzing electronic mails. ContactFinder [9] 10] is an agent style system that monitors discussion groups and uses various heuristics to extract personal contacts and identify specific areas. 12] builds the expertise index by mining relationships ....

....the expert finding problem from a different perspective [5] 6] They postulate that the best way of finding an expert is through what is called referral chaining . The referral chain was built by mining the public web pages. The information source of previous systems can be electronic mails [15], discussion groups [9] 10] personal web pages [5] 6] web browsing patterns [13] 2] various documents reports related to particular users [18] 19] 12] etc. Most of the previous systems use textual analysis as the basis for expert expertise profiling except for [15] which uses graph analysis to ....

[Article contains additional citation context not shown here]

M. F. Schwartz and D. M. Wood, Discovering Shared Interests Using Graph Analysis, Communications of the ACM, vol. 36, no. 8, pp. 78 -- 89, 1993.

....networks have been the object of much research. One classic paper is that by Milgram [20] which estimated that every person in the world is only six acquaintances away from every other. Some recent social network research uses the Internet as a source of data. For instance, Schwartz and Wood [23] mined social relationships from email logs, the ReferralWeb project mined a social network from a wide variety of publicly available online information [14] and the COBOT project gathered social statistics from participant interactions in the LambdaMoo MUD [11] Our network was mined from a ....

M. F. Schwartz and D. C. M. Wood. Discovering shared interests using graph analysis. Communications of the ACM, 36(8):78-80, 1993.

.... reports [FD92, FC97] Usenet news [FS91, Bac91, SK92, Ste92b, Bac92, JH92, SM93, Mae94, KHL 94, RIS 94, MS94, Lan95, YG95, Moc96, MRK97, MRK 97] electronic mail [Mye80, MGT 87, Pol88, GNOT92, Ter91, Ter93, LMM94] books [Ric79b, MR98] application program know how [LN98] the finding of experts [SW93, KSS96] Web pages [RM96, HT96, Bal97, THA 97, PB97, RP97, Bie98] classified ads [GGKS95] movies [Kay95, HRF95, AKK97, AKK98] music [Sha94, SM95, Loe92] and TV [EHWS96, DtHh98, Bau96] In Chapter 2, we will discuss some of them in more detail. Figure 2a shows the general architecture of IF ....

M.F. Schwartz and D.C.M. Wood. Discovering shared interests using graph analysis. Communications of the ACM, 36(8):78-89, August 1993.

....Their work is based on the observation that expert finding is an inherently complex process. Hence, they postulate that the best way of finding an expert is through what is called referral chaining whereby a seeker finds the needed expert through referral by colleagues. Schwartz and Wood [26] have also tried to enable locating people with related interests or expertise by analyzing the graph which they called specialization subgraph formed by email communication patterns (rather than their contents) They use a set of heuristic algorithms to uncover shared interest relationships ....

M. F. Schwartz and D. M. Wood, Discovering Shared Interests Using Graph Analysis,Communications of the ACM, vol. 36, no. 8, pp. 78 --- 89, 1993.

....that follow scenarios. Furthermore, these activities make and characterize relationships within and between societies, streams, and structures. Each activity happens in a setting, arena, or space. The relationships developed can be seen under the optics of larger structures (e.g. social networks [102, 57]) 5.1.3 Components: What constitutes DLs Digital libraries can contain repositories of knowledge, information, data, metadata, relationships, logs, annotations, user profiles, and documents, all which can be interpreted as distinct forms of digital objects, according to their particular ....

M. F. Schwartz and D. C. M. Wood. Discovering shared interests using graph analysis. Communications of the ACM, 36(8):78, Aug. 1993.

....might be applicable in our context. In retrospect, the earliest sign of the potential of viral marketing was perhaps the classic paper by Milgram [31] estimating that every person in the world is only six edges away from every other, if an edge between i and j means i knows j. Schwartz and Wood [37] mined social relationships from email logs. The ReferralWeb project mined a social network from a wide variety of publicly available online information [24] and used it to help individuals nd experts who could answer their questions. The COBOT project Free Movie 0 2 4 6 8 10 12 1 1.5 2 ....

M. F. Schwartz and D. C. M. Wood. Discovering shared interests using graph analysis. Communications of the ACM, 36(8):78-89, 1993.

....are linked by a chain of six or fewer first name acquaintances. Researchers have replicated this phenomenon in many contexts, ranging from groups of workers within a corporation (Wasserman and Galaskiewicz 1994; Krackhardt and Hanson 1993) to patterns of e mail communication around the world (Schwartz and Wood 1993). The idea that everyone is connected this way has been popularized by the phrase six degrees of separation as well as by games based on tracing the networks around celebrities. A well known pastime in the mathematics community is determining an individual s Erdos number, which is the number of ....

....between the individuals. There are many possible sources for determining direct relationships. At one extreme, which we reject as too burdensome, users could be required to enter lists of close colleagues. Analysis of e mail logs provides a rich source of relationships, as shown by Schwartz and Wood (1993). In fact, the initial version of our system derived its network by analyzing mail archives (Kautz, Selman, and Milewski 1996) However, the use of such information raises concerns of Articles SUMMER 1997 31 Figure 1. Success Rate as a Function of Responsiveness and Referral Accuracy. 0 ....

Schwartz, M. F., and Wood, D. C. M. 1993. Discovering Shared Interests Using Graph Analysis. Communications of the ACM 36(8): 78--89.

....Kautz, et al. 1997) Their work is based on the observation that expert finding is an inherently complex process. Hence, they postulate that the best way of finding an expert is through what is called referral chaining whereby a seeker finds the needed 12 expert through referral by colleagues. Schwartz and Wood (1993) have also tried to enable locating people with related interests or expertise by analyzing the graph which they called specialization subgraph formed by email communication patterns (rather than their contents) They use a set of heuristic algorithms to uncover shared interest relationships ....

Schwartz, M. F., & Wood, D. M. (1993). Discovering Shared Interests Using Graph Analysis. Communications of the ACM. 36 (8): 78 --- 89.

....effort. One method for collaborator discovery is known as referral chaining. Referral chaining is a form of link analysis that exploits the fact that people with common interests tend to congregate within the same social circles. Techniques which automate this approach can be found in Schwartz (Schwartz and Wood 1993) in which shared interests are identified based on analysis of email traffic patterns, and Kautz (Kautz, Selman, and Mehul 1997) in which shared interests are identified from co occurrences of names in public documents. There is a problem with these approaches, however, for our purposes. The ....

Schwartz, M.; and Wood, D. 1993. Discovering Shared Interests Using Graph Analysis. In Communications of the ACM, 36(8): 78-89.

....concepts and grouped accordingly. A similar process is followed in the internet resource discovery method discussed in [23] In this case, however, the derivation of areas of interest (global concepts) are based on a list of topics which are searched for automatically in e mail messages (see [26]) It is believed that organization of the UoD should follow as a natural consequence of the interaction of brokers. Broker interactions should result in the dynamic formation (and re formation) of node clusters of mutual interest, based upon pre established subdomains. For example, if a broker ....

M.F. Schwartz, D.C.M. Wood, "Discovering Shared Interests using Graph Analysis", Communications of the ACM, 1993, vol.36, no.8, pp. 78-89.

....the individuals has been discovered. There are many possible sources for determining direct relationships. At one extreme, which we reject as too burdensome, users could be required to enter lists of close colleagues. Analysis of email logs provides a rich source of relationships, as shown by Schwartz and Wood (1993). In fact, the initial version of our system derived its network by analyzing mail archives (Kautz, Selman, and Milewski, 1996) However, the use of such information raises concerns of privacy and security that are hard to allay. The current ReferralWeb system uses the co occurrence of names in ....

Schwartz, M. F. and Wood, D. C. M. Discovering shared interests using graph analysis. Comm. ACM 36, 8 (1993), 78--89.

....all ten minute periods for one month. Animating these, as described earlier, gives an excellent dynamic view of the network, and has led to some interesting insights. 7] 5.2 Email Communications Within the technical community, electronic mail is widely used for interpersonal communications. [8] In our location, members of the technical staff receive 30 to 40 messages per day. For nearly a year we logged the sender, receiver, message size, and time for every email message sent or received by members of one of our departments at AT T Bell Laboratories, who volunteered for this study. ....

Michael F. Schwartz and David C. M. Wood, "Discovering Shared Interests Using Graph Analysis," Communications of the ACM, Vol. 36(8), 1993, pp. 78-89.

....of a single parameter used in the simulation experiment for Table 1. See the caption of each table. The results in these tables show again that one can trade referral accuracy for responsiveness, and vice versa. As we noted above, the simulation experiments were run on random graphs. Schwartz and Wood (1993) have created and analyzed graphs of email communication between groups of people, based on email logs they obtained in the late 1980 s from system administrators around the world. An interesting direction for future work would be to run our simulations on graphs patterned after such actual ....

Schwartz, M. F. and Wood, D. C. M. (1993). Discovering shared interests using graph analysis. Comm. ACM, 36(8), 1993, 78--89.

....Contact: AT T Bell Laboratories Rm 1G 351, 1000 East Warrenville Road, Naperville, IL 60566, email: eick research.att.com network. 1 The clutter comes from the long lines connecting distant pairs of nodes that cause overplotting. For examples of particularly complex graph displays see [SW93]. Figure 1: World wide internet traffic over a two hour period, with the color and thickness of the lines encoding the traffic. There are several possible solutions to the display clutter problem: ffl using curves, perhaps spline based, instead of straight lines to connect distant nodes ....

Michael F. Schwartz and David C. M. Wood. Discovering shared interests using graph analysis. Communications of the ACM, 36(8):78--89, 1993.

....based on frequency of cooccurrence of names in the entire WWW. The spiders perform selective web crawling to generate lists of potential pairs, and then query full text Web indexes such as Altavista to gather precise statistics. ffl Email logs provide a very good indicator of relatedness (Schwartz and Wood 1993). However, because of privacy concerns we do not use such records. In fact, our experiments with an early version of our system that did use email information (Kautz, Selman, and Milewski 1993) ended when the test subjects refused to allow any program to analyze their email, despite the elaborate ....

Schwartz, M. F. and Wood, D. C. M. (1993). Discovering shared interests using graph analysis. Comm.

No context found.

Schwartz, M., Wood, D.: Discovering shared interests using graph analysis. Communications of the ACM 36 (1993) 78--89

No context found.

M. F. Schwartz and D. C. M. Wood. Discovering Shared Interests Using Graph Analysis. Communications of the ACM, Vol. 36(8):pages 78--89, August 1993.

No context found.

M.F. Schwartz and D.C.M. Wood. "Discovering Shared Interests Using Graph Analysis ". Communications of the ACM, Vol. 36(8):pages 78--89, August 1993.

No context found.

M. F. Schwartz and D. C. M. Wood. Discovering shared interests using graph analysis. Communications of the

No context found.

M.F. Schwartz and D.C.M. Wood. "Discovering Shared Interests Using Graph Analysis ". Communications of the ACM, Vol. 36(8):pages 78--89, August 1993.

No context found.

Schwartz,F. M. & Wood, M. C. (1993). Discovering Shared Interests Using Graph Analysis, Communications of ACM, no.36, vol.8, pp.78-89.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC