Limited Offer Get 25% off — use code BESTW25
No AI No Plagiarism On-Time Delivery Free Revisions
Claim Now
Prices start at $5 per page · plagiarism-free papers · free revisions for 14 days 24/7 support: support@bestwriters.com
Uncategorized

Faculty of Computer Science

  • Posted
  • Authoradmin
  • Length2,302 words
  • Reading time11 min
Write My Assignment FREE

Weighted PageRank Algorithm
Wenpu Xing and Ali Ghorbani
Faculty of Computer Science
University of New Brunswick
Fredericton, NB, E3B 5A3, Canada
E-mail: m0yac,ghorbani@unb.ca
Abstract
With the rapid growth of the Web, users get easily lost in
the rich hyper structure. Providing relevant information to
the users to cater to their needs is the primary goal of website owners. Therefore, finding the content of the Web and
retrieving the users’ interests and needs from their behavior have become increasingly important. Web mining is used
to categorize users and pages by analyzing the users’ behavior, the content of the pages, and the order of the URLs
that tend to be accessed in order. Web structure mining
plays an important role in this approach. Two page ranking algorithms, HITS and PageRank, are commonly used
in web structure mining. Both algorithms treat all links
equally when distributing rank scores. Several algorithms
have been developed to improve the performance of these
methods. The Weighted PageRank algorithm (WPR), an extension to the standard PageRank algorithm, is introduced
in this paper. WPR takes into account the importance of
both the inlinks and the outlinks of the pages and distributes
rank scores based on the popularity of the pages. The results
of our simulation studies show that WPR performs better
than the conventional PageRank algorithm in terms of returning larger number of relevant pages to a given query.
Keywords: Web Mining, Web Structure Mining, HITS,
PageRank, Weighted PageRank
1. Introduction
In the highly competitive world and with the broad use
of the Web in e-commerce, e-learning, and e-news, finding
users’ needs and providing useful information are the primary goals of website owners. Therefore, analyzing users’
patterns of behavior becomes increasingly important.
Web mining is used to discover the content of the Web,
the users’ behavior in the past, and the webpages that the
users want to view in the future. Web mining consists
of Web Content Mining (WCM), Web Structure Mining
(WSM), and Web Usage Mining (WUM) [6, 7, 9]. WCM
deals with the discovery of useful information from web
content. WSM discovers relationships between web pages
by analyzing web structures. WUM ascertains user profiles
and the users’ behavior recorded inside the web logfile.
WCM and WUM have been studied by many researchers
who have achieved valuable results. Based on the topology
of the hyperlinks, WSM categorizes web pages and generates related patterns, such as the similarity and the relationships between different Web sites. Technically, WCM
focuses mainly on the structure within a document (the
inner-document level) while WSM tries to discover the link
structure of the hyperlinks between documents (the interdocument level). The numbers of inlinks (links to a page)
and of outlinks (links from a page) are valuable information
in web mining. This is due to the facts that a popular webpage is often referred to by other pages and that an “important” webpage contains a high number of outlinks. Therefore, WSM is seen as an important approach to web mining.
This paper focuses on WSM and provides a new Weighted
PageRank Algorithm.
The rest of this paper is organized as follows. A brief
background review of web structure mining is presented
in the next section. Section 3 presents the PageRank algorithm, a commonly used algorithm in WSM. An extended PageRank algorithm called the Weighted PageRank
algorithm (WPR) is described in Section 4. Different components involved in the implementation and evaluation of
WPR are presented in Section 5. The experimental results
and their implication for WPR are given in Section 6. Section 7 summarizes the conclusions of the present study. Finally, the result sets of PageRank and WPR for the query
”travel agent” are given in Appendices A and B respectively.
2. Background
With the rapid growth of the Web, providing relevant
pages of the highest quality to the users based on their
queries becomes increasingly difficult. The reasons are that
some web pages are not self-descriptive and that some links
exist purely for navigational purposes. Therefore, finding
appropriate pages through a search engine that relies on web
contents or makes use of hyperlink information is very difficult.
To address the problems mentioned above, several algorithms have been proposed. Among them are PageRank [10]
and Hypertext Induced Topic Selection (HITS) [2, 9] algorithms. PageRank is a commonly used algorithm in Web
Structure Mining. It measures the importance of the pages
by analyzing the links [1, 8]. PageRank has been developed by Google and is named after Larry Page, Google’s
co-founder and president[10]. PageRank ranks pages based
on the web structure.
Google first retrieves a list of relevant pages to a given
query based on factors such as title tags and keywords. Then
it uses PageRank to adjust the results so that more “important” pages are provided at the top of the page list [10]. The
Pagerank algorithm is described in detail in the next section.
HITS ranks webpages by analyzing their inlinks and outlinks. In this algorithm, webpages pointed to by many hyperlinks are called authorities whereas webpages that point
to many hyperlinks are called hubs [4, 5, 11]. Authorities
and hubs are illustrated in Figure 1.
Figure 1. Hubs and authorities
Hubs and authorities are assigned respective scores.
Scores are computed in a mutually reinforcing way: an authority pointed to by several highly scored hubs should
be a strong authority while a hub that points to several highly scored authorities should be a popular hub
[4, 5]. Let ap and hp represent the authority and hub scores
of page p, respectively. B(p) and I(p) denote the set of referrer and reference pages of page p, respectively. The
scores of hubs and authorities are calculated as follows [2, 4, 5]:
a
p = X
q2B(p)
h
q (1)
h
p = X
q2I(p)
a
q (2)
Figure 2 shows an example of the calculation of authority and hub scores.
Figure 2. An example of HITS operations
HITS is a purely link-based algorithm. It is used to rank
pages that are retrieved from the Web, based on their textual contents to a given query. Once these pages have been
assembled, the HITS algorithm ignores textual content and
focuses itself on the structure of the Web only. Some difficulties arise from this feature [2]:
• HITS frequently returns more general webpages on
an otherwise narrowly focused topic because the web
does not contain many resources for the topic,
• Topic drift occurs while the hub has multiple topics because all of the outlinks of a hub page get equivalent
weights, and
• Some popular sites that are not highly relevant to the
given query gain overhead weight values.
The CLEVER algorithm is an extension of standard
HITS and provides an appropriate solution to the problems that result from standard HITS [2]. CLEVER assigns
a weight to each link based on the terms of the queries and
end-points of the link. It combines anchor text to set weights
to the links as well. Moreover, it breaks large hub pages into
smaller units so that each hub page is focused on as a single topic. Finally, in the case of a large number of pages
from a single domain, it scales down the weights of pages
to reduce the probabilities of overhead weights [2].
Another major shortcoming of standard HITS is that
it assumes that all links pointing to a page are of equal
weight and fails to recognize that some links might be
more important than others. A Probabilistic analogue of
the HITS Algorithm(PHITS) has been developed to solve
this problem[3]. PHITS provides a probabilistic interpretation of term-document relationships and identifies authoritative documents. In the experiment on a set of hyperlinked
documents, PHITS demonstrates better results compared to
those obtained by standard HITS. The most important feature of the PHITS algorithm is its ability to estimate the actual probabilities of authorities compared to the scalar magnitudes of authority that are provided by standard HITS[3].
3. The PageRank Algorithm
The PageRank algorithm, one of the most widely used
page ranking algorithms, states that if a page has important
links to it, its links to other pages also become important.
Therefore, PageRank takes the backlinks into account and
propagates the ranking through links: a page has a high rank
if the sum of the ranks of its backlinks is high [8, 10]. Figure 3 shows an example of backlinks: page A is a backlink
of page B and page C while page B and page C are backlinks of page D.
Figure 3. An example of backlinks
3.1. Simplified PageRank
A slightly simplified version of PageRank is defined as
[8]:
PR(u) = c X
v2B(u)
PR(v)
Nv
(3)
where u represents a web page. B(u) is the set of pages that
point to u. PR(u) and PR(v) are rank scores of page u and
v, respectively. Nv denotes the number of outgoing links of
page v. c is a factor used for normalization. Figure 4 shows
an example in which c = 1.0 to simplify the calculation.
In PageRank, the rank score of a page, p, is evenly divided among its outgoing links. The values assigned to the
outgoing links of page p are in turn used to calculate the
Figure 4. An example of simplified version of
PageRank
ranks of the pages to which page p is pointing. The rank
scores of pages of a website could be calculated iteratively
starting from any webpage. Within a website, two or more
pages might connect to each other to form a loop. If these
pages did not refer to but are referred to by other webpages
outside the loop, they would accumulate rank but never distribute any rank. This scenario is called a rank sink [8].
3.2. PageRank
To solve the rank sink problem, we observed the users’
activities. A phenomenon is found that not all users follow
the existing links. For example, after viewing page a, some
users may not decide to follow the existing links but directly go to page b, which is not directly linked to page a.
For this purpose, the users just type the URL of page b into
the URL text field and jump to page b directly. In this case,
the rank of page b should be affected by page a even though
these two pages are not directly connected. Therefore, there
is no absolute rank sink.
Based on the consideration of the phenomenon mentioned above, the original PageRank is published [8, 10]:
PR(u) = (1 ! d) + d X
v2B(u)
PR(v)
Nv
(4)
where d is a dampening factor that is usually set to 0.85.
We also could think of d as the probability of users’ following the links and could regard (1 ! d) as the pagerank distribution from non-directly linked pages.
To test the utility of the PageRank algorithm, Google
applied it to the Google search engine [8]. In the experiments, the PageRank algorithm works efficiently and effectively because the rank value converges to a reasonable tolerance in the roughly logarithmic (log n) [8, 10].
The rank score of a web page is divided evenly over the
pages to which it links. Even though the PageRank algorithm is used successfully in Google, one problem still ex
ists: in the actual web, some links in a web page may be
more important than are the others.
4. Weighted PageRank (WPR)
The more popular webpages are, the more linkages that
other webpages tend to have to them or are linked to
by them. The proposed extended PageRank algorithm–a
Weighted PageRank Algorithm–assigns larger rank values
to more important (popular) pages instead of dividing the
rank value of a page evenly among its outlink pages. Each
outlink page gets a value proportional to its popularity (its
number of inlinks and outlinks). The popularity from the
number of inlinks and outlinks is recorded as W in
(v,u) and
W out
(v,u), respectively.
W in
(v,u) is the weight of link(v, u) calculated based on
the number of inlinks of page u and the number of inlinks
of all reference pages of page v.

W in
(v,u) =
Iu
Pp2R(v) Ip
(5)

where Iu and Ip represent the number of inlinks of page u
and page p, respectively. R(v) denotes the reference page
list of page v.
W out
(v,u) is the weight of link(v, u) calculated based on
the number of outlinks of page u and the number of outlinks of all reference pages of page v.

W out
(v,u) =
Ou
Pp2R(v) Op
(6)

where Ou and Op represent the number of outlinks of page
u and page p, respectively. R(v) denotes the reference page
list of page v.
Figure 5 shows an example of some links of a hypothetical website.
Figure 5. Links of a website
In this example, Page A has two reference pages: p1 and
p2. The inlinks and outlinks of these two pages are Ip1 = 2,
Ip
2 = 1, Op1 = 2, and Op2 = 3. Therefore,
W in
(A,p1) = Ip1/(Ip1 + Ip2) = 2 3
and
W out
(A,p1) = Op1/(Op1 + Op2) = 2 5
Considering the importance of pages, the original PageRank formula is modified as
PR(u) = (1 ! d) + d X
v2B(u)
PR(v)W(in v,u)W(out v,u) (7)
5. Experiments
To evaluate the WPR algorithm, we implemented WPR
and the standard PageRank algorithms to compare their results. Figure 6 illustrates different components involved in
the implementation and evaluation of the WPR algorithm.
The simulation studies we have carried out in this work
consist of six major activities:
1. Finding a web site: Finding a web site with rich hyperlinks is necessary because the standard PageRank and
the WPR algorithms rely on the web structure. After
comparing the structures of several web sites, the website of Saint Thomas University, in Fredericton, has
been chosen.
2. Building a web map: There is no web map available for
this website. A free spider software–JSpider–is used to
generate the required web map.
3. Finding the root set: A set of pages relevant to a given
query is retrieved using the IR search engine embedded in the web site. This set of pages is called the root
set.
4. Finding the base set: A base set is created by expanding the root set with pages that directly point to or are
pointed to by the pages in the root set.
5. Applying algorithms: The Standard PageRank and the
WPR algorithms are applied to the base set.
6. Evaluating the results: The algorithms are evaluated by
comparing their results.
Normally, websites in different domains focus on different topics. Usually, the websites have rich linkages to describe the focused topics. On the other hand, they do a poor
job describing non-focused topics. For example, the websites of most universities have a lot of information about
scholarships and courses whereas the websites of travel
companies mainly provide travel paths and scenes around

The post Faculty of Computer Science appeared first on My Assignment Online.

Plagiarism Free Assignment Help

Expert Help With This Assignment — On Your Terms

✓ Native UK, USA & Australia writers ✓ Deadline from 3 hours ✓ 100% Plagiarism-Free — Turnitin included ✓ Unlimited free revisions ✓ Free to submit — compare quotes

admin

Academic writer and editor at BestWriters.

Get your paper off your plate tonight

Post your instructions for free. Prices start at $5 per page.

Place an order