Bernado A. Huberman and Lada A. Adamic. Evolutionary dynamics of the world wide web. Technical report, Xerox Palo Alto Research Center, February 1999. http://www.hpl.hp.com/research/ idl/papers/webgrowth/.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....change dynamics of documents. Cho et al. [9] computed the lifespan of pages in five different domains, namely .gov, net, org, edu, and .com, and showed that it varies widely. Smaller studies on how often web pages change were performed by Wills et al. [19] and Douglis et al. [11] Huberman et al. [1] presented a theory for the growth dynamic of the Web that takes into account the growth rates in the number of pages per site, as well as the fact that new sites are created at different times. Brewington [7] developed a different model of web page changes. In the area of web graph analysis, ....

L. Adamic and B. Huberman. Evolutionary dynamics of the world wide web. Nature, September 1999.

....have arisen as a possible alternative to power law distributions across many elds. 1 Introduction Power law distributions (also often referred to as heavy tail distributions, Pareto distributions, Zip an distributions, etc. are now pervasive in computer science; see, for example, [1, 8, 7, 9, 16, 19, 21, 22, 24, 25, 27, 28, 33, 34, 40, 41, 43, 45, 47, 50, 61, 69]. This paper was speci cally motivated by a recent paper by Downey [25] challenging the now conventional wisdom that le sizes are governed by a power law distribution. The argument was substantiated both by collected data and by the development of an underlying generative model which ....

....rate in a perfect market. Other applications in for example geology and atmospheric chemistry are given in [23] More recently, as described below, Adamic and Huberman suggest that multiplicative processes may describe the growth of sites on the Web as well as the growth of user trac on Web sites [33, 34]. Lognormal distributions have also been suggested for le sizes [8, 9, 25] The connection between multiplicative processes and the lognormal distribution can be traced back to Gibrat around 1930 [30, 31] although Kapteyn described in other terms an equivalent process in 1903 [38] and ....

[Article contains additional citation context not shown here]

B. A. Huberman and L. A. Adamic. Evolutionary dynamics of the World Wide Web. Technical Report, Xerox Palo Alto Research Center, 1999.

....surprising that lognormal distributions arise as a possible alternative to power law distributions. 1 Introduction Power law distributions (also often referred to as heavy tail distributions, Pareto distributions, Zip an distributions, etc. are now pervasive in computer science; see, e.g. [6, 8, 7, 13, 16, 18, 19, 21, 22, 24, 28, 29, 34, 35, 37, 38, 39, 50, 56]. This paper was motivated by a recent paper by Downey [22] challenging the now conventional wisdom that le sizes are governed by a power law distribution. The argument was substantiated both by collected data and by the development of an underlying generative model which suggested that le ....

....prevailing interest rate in a perfect market. Other applications in for example geology and atmospheric examples are given in [20] More recently, Adamic and Huberman suggest that multiplicative processes may describe the growth of links on the Web as well as the growth of user trac on Web sites [28, 29], and lognormal distributions have been suggested for le sizes [8, 7, 22] The connection between multiplicative processes and the lognormal distribution can be traced back to Gibrat around 1930 [26, 27] although Kapteyn [32] described in other terms an equivalent process in 1903, and McAlister ....

[Article contains additional citation context not shown here]

B. A. Huberman and L. A. Adamic. Evolutionary Dynamics of the World Wide Web. Technical Report, Xerox Palo Alto Research Center, 1999.

....due to the fact that outlinks are often created for reasons other than preferential attachment, for example, in order to maintain the local structure of a Web site. Another interpretation of i is the number of pages within Web sites (referred to as webpages) In this case, Huberman and Adamic [HA99] reported a power law distribution with exponent 1.85, derived from a 250,000 Web site crawl. Our model cannot explain this observation as the exponent is less than two. A more recent result from a private communication with Adamic reported an exponent of 2.2, derived from a 1.6 million Web site ....

B.A. Huberman and L.A. Adamic. Evolutionary dynamics of the World Wide Web. Nature, 399:131, 1999.

.... = 0 and = 1. In this case, for x 1, f(x) and for x 1, 1 . A key characteristic of the double Pareto distribution is that it has a power law at both tails. That is, if we look at the cumulative distribution function (cdf) on a log log plot, it Huberman and Adamic [15, 16] also examine this distribution and conclude that it has a power law distribution. Their earlier work, however, fails to note the behavior of the distribution goes through a phase shift, which Reed clari es. 5 will also have a linear tail (for the small les) This provides a test for seeing ....

B. A. Huberman and L. A. Adamic. Evolutionary Dynamics of the World Wide Web. Technical Report, Xerox Palo Alto Research Center, 1999.

....[LG99] The 11 months that had passed showed a significant increase in the number of indexable documents found. The lower bound of size was estimated to 800 million documents, and the search engines had significant less coverage of the indexable web. The maximum coverage estimated was 16 . In [HA99], a theory for the growth dynamics of the World Wide Web is presented. Two stochastic classes are considered, namely the growth rates of pages per site, and the growth of new sites. A universal power law is predicted for the distribution of the number of pages per site. The paper brings theories ....

Bernardo A. Huberman and Lada A. Adamic. Evolutionary Dynamics of the World Wide Web. Technical Report. Xerox Palo Alto Research Center. February 1999.

....surprising that lognormal distributions arise as a possible alternative to power law distributions. 1 Introduction Power law distributions (also often referred to as heavy tail distributions, Pareto distributions, Zipfian distributions, etc. are now pervasive in computer science; see, e.g. [6, 8, 7, 13, 16, 18, 19, 21, 22, 24, 28, 29, 34, 35, 37, 38, 39, 49, 55]. 1 This paper was motivated by a recent paper by Downey [22] challenging the now conventional wisdom that file sizes are governed by a power law distribution. The argument was substantiated both by collected data and by the development of an underlying generative model which suggested that file ....

....prevailing interest rate in a perfect market. Other applications in for example geology and atmospheric examples are given in [20] More recently, Adamic and Huberman suggest that multiplicative processes may describe the growth of links on the Web as well as the growth of user tra#c on Web sites [28, 29], and lognormal distributions have been suggested for file sizes [8, 7, 22] The connection between multiplicative processes and the lognormal distribution can be traced back to Gibrat around 1930 [26, 27] although Kapteyn [32] described in other terms an equivalent process in 1903, and ....

[Article contains additional citation context not shown here]

B. A. Huberman and L. A. Adamic. Evolutionary Dynamics of the World Wide Web. Technical Report, Xerox Palo Alto Research Center,

....we have b = ln x) 2 2# 2 . Because of this, there are two di#erent behaviors, depending on whether x # 1 or x # 1. Let C 1 = # # ## ( #) 2 2# # and let C 2 = # 2 ## ( #) 2 2# # #. For x # 1, f(x) C 1 x 1 C 2 , so the result is a power 2 Huberman and Adamic [15, 16] also examine this distribution and conclude that it has a power law distribution. Their earlier work, however, fails to note the behavior of the distribution goes through a phase shift, which Reed clarifies. 5 law distribution. For x # 1, f(x) C 1 x 1 C2 . In particular, a case we use ....

B. A. Huberman and L. A. Adamic. Evolutionary Dynamics of the World Wide Web. Technical Report, Xerox Palo Alto Research Center,

....exponent of 2.72 we would have to assume # # 3.42. However, Simon s original model would predict an exponent of about 2.16 for outlinks, similar to that for inlinks. Another interpretation of i is the number of pages within Web sites (referred to as webpages) In this case, Huberman and Adamic [HA99] reported a power law distribution with exponent 1.85, derived from a 250,000 Web site crawl. Our model cannot explain this observation as the exponent is less than two. A more recent result from a private communication with Adamic reported an exponent of 2.2, derived from a 1.6 million Web site ....

B.A. Huberman and L.A. Adamic. Evolutionary dynamics of the World Wide Web. Nature, 399:131, 1999.

....Others, e.g. Henzinger et al. 1999; Henzinger et al. 2000) have attempted to achieve uniform web page sampling using random walks on the Web link graph. Characteristics of the Web graph structure, growth rate and server page size distribution have been investigated by many authors, including (Huberman and Adamic, 1999; Lawrence and Giles, 1999; Faloutsos et al. 1999; Broder et al. 2000) Various attempts have been made to measure the retrieval performance of commercial search engines, e.g. Gordon and Pathak, 1999; Hawking et al. 2001) These studies treat the whole of the web as a document collection and ....

....was computed in an analogous way. A server server sparse matrix recorded the number of links between each pair of servers. 5.4 Phase 4: Selection of WT10g servers 5.4.1 Size representativeness A representative distribution of server sizes was a very important goal for WT10g. Huberman and Adamic (Huberman and Adamic, 1999) approximate the distribution of Web server sizes using the power law P (n s ) Cn Gammafi s where P (n s ) is the probability that a Web site has n s pages and C and fi 1 are constants. This means that there are a few very large servers, and many very small ones. 7 0.01 0.1 1 10 100 ....

Huberman, B. A. and Adamic, L. A. (1999). Evolutionary Dynamics of the World Wide Web. http://www.parc.xerox.com/istl/groups/iea/www/growth.html.

....Others, e.g. Henzinger et al. 1999; Henzinger et al. 2000) have attempted to achieve uniform web page sampling using random walks on the Web link graph. Characteristics of the Web graph structure, growth rate and server page size distribution have been investigated by many authors, including (Huberman and Adamic, 1999; Lawrence and Giles, 1999; Faloutsos et al. 1999; Broder et al. 2000) Various attempts have been made to measure the retrieval performance of commercial search engines, e.g. Gordon and Pathak, 1999; Hawking et al. 2001) These studies treat the whole 1 At the time of writing, more than 40 ....

....was computed in an analogous way. A serverserver sparse matrix recorded the number of links between each pair of servers. 5.4 Phase 4: Selection of WT10g servers 5.4.1 Size representativeness A representative distribution of server sizes was a very important goal for WT10g. Huberman and Adamic (Huberman and Adamic, 1999) approximate the distribution of Web server sizes using the power law P (n s ) Cn s where P (n s ) is the probability that a Web site has n s pages and C and 1 are constants. This means that there are a few very large servers, and many very small ones. 7 0.01 0.1 1 10 100 1 2 4 8 ....

Huberman, B. A. and Adamic, L. A. (1999). Evolutionary Dynamics of the World Wide Web. http://www.parc.xerox.com/istl/groups/iea/www/growth.html.

....using power law topologies. I. Introduction Recent studies have shown that Internet graphs follow power laws [1] i.e. certain graph metrics follow the distribution y x ff . This phenomenon has been observed in router topology, inter domain topology [1] and the worldwide web [2] 3] [4], 5] 6] Previous metrics used to characterize Internet graphs have focussed on averages: for example, average out degree of routers. While these metrics are important, they do not capture higher order properties such as the power laws. Network studies often simulate artificial network ....

B. A. Huberman and L. A. Adamic, "Evolutionary dynamics of the world wide web," Tech. Rep., Xerox Palo Alto Research Center, February 1999.

....the impact of power law topologies. 1 Introduction Recent studies have shown that Internet graphs follow power laws [5] i.e. certain graph metrics follow the distribution y x ff . This phenomenon has been observed in router topology, inter domain topology [5] and the world wide web [1, 2, 6, 7, 8]. Previous metrics used to characterize Internet graphs have focussed on averages: for example, average out degree of routers. While these metrics are important, they do not capture higher order properties such as the power laws. Network studies often simulate artificial network topologies to ....

B. A. Huberman and L. A. Adamic. Evolutionary dynamics of the world wide web. Technical report, Xerox Palo Alto Research Center, February 1999.

No context found.

Bernado A. Huberman and Lada A. Adamic. Evolutionary dynamics of the world wide web. Technical report, Xerox Palo Alto Research Center, February 1999. http://www.hpl.hp.com/research/ idl/papers/webgrowth/.

No context found.

HUBERMAN, B., AND ADAMIC, L. Evolutionary dynamics of the world wide web. Tech. rep., Xerox Palo Alto Research Center, 1999. 16

No context found.

Bernardo A. Huberman and Lada A. Adamic. Evolutionary dynamics of the World Wide Web. Condensed Matter, January 1999. (paper 9901071).

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