DBMS REVIEW-3

G.BALAVIGNESH-10MSE1072
HARSHAVARDHAN-10MSE1077
ABSTRACT
This paper explores different Page Rank
algorithms like Page Rank (PR), WPR
(Weighted Page Rank), HITS
(Hyperlink-Induced Topic Search),
weighted content page rank algorithms
are discussed and compared.
Google Architecture
PAGE RANK?
Page ranking algorithms are used by
the search engines to present the
search results by considering the
relevance, importance and content
score and web mining techniques to
order them according to the user
interest
Expanded Definition
R(u): page rank of page u
c: factor used for normalization (<1)
B
u
: set of pages pointing to u
N
v
: outbound links of v
R(v): page rank of site v that points to u
E(u): distribution of web pages that a random
surfer periodically jumps (set to 0.15)
) (
) (
) ( u cE
N
v R
c u R
u
B v
v
+ =

e
Weighted Page Rank
Extended PageRank algorithm- Weighted
PageRank Assigns large rank value to more
important 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 (no. of inlinks
and outlinks). The popularity from the
number of inlinks and outlinks and is
recorded as Win(V,U) and Wout(V,U) In
Weighted PageRank all links are not equally
distributed i.e. unequal distribution
Formula


Where d = damping factor to set value
0 to 1
Win(U,V)=weighted of link (U,V)

Inlinks


Where IU = number of inlinks of
page u , Ip = number of inlinks of
page p
R(v)=Reference page text of page ,
Wout(V,U)= weight of outlink(V,U)

Outlinks



Where Ou=number of outlink of
page u , Op= number of outlink of
page p
Limitations of Pagerank and
Weighted Pagerank
PAGE RANK
PageRank is equally distributed to
outgoing links.
It is purely based on the number of
inlinks and outlinks.

Weighted PageRank

Some pages may be irrelevant to a
given query, it still receives the highest
rank because it has many inlinks and
many outlinks.
There is a less determination of the
relevancy of the pages to a given query
This algorithm relies mainly on the
number connected inlinks and outlinks
Proposed Weighted Page
Content Rank
Get the required relevant documents
easily on the top few pages.
Employs Web content mining-
Mining, extraction and integration of
useful data, information and knowledge
from Web page contents.
Employs Web Structure mining
- Graph theory to analyze the node and
connection structure of a web site.
- Extracting patterns from hyperlinks in
the web and mining the document
structure.

Modified architecture


where PR(U)=PageRank of page U,
B(U)= Set of all pages referring to
page U.
D= Damping factor between 0 and 1,
Cw=Content weight of page U
Pw=Probability weight of page U


Basic Idea
Back-links coming from important pages
convey more importance to a page. For
example, if a web page has a link off the
yahoo home page, it may be just one link but
it is a very important one.
A page has high rank if the sum of the ranks
of its back-links is high. This covers both the
case when a page has many back-links and
when a page has a few highly ranked back-
links.
Definition
My pages rank is equal to the sum of
all the pages pointing to me.
v f rom links of number N
u to links with pages of set B
N
v Rank
u Rank
v
u
B v
v
u
=
=
=

e
) (
) (
Simplified PageRank Example
Rank(u) = Rank of
page u , where c is
a normalization
constant (c < 1 to
cover for pages with
no outgoing links).
Problem 1 - Rank Sink
Page cycles pointed by some incoming link.





Loop will accumulate rank but never
distribute it.
Problem 2 - Dangling Links
In general, many Web pages do not have either back links or forward
links.











Dangling links do not affect the ranking of any other page directly, so
they are removed until all the PageRanks are calculated.
Random Surfer Model
PageRank corresponds to the probability
distribution of a random walk on the web
graphs.




Solution Escape Term
Escape term: E(u) can be thought of as the
random surfer gets bored periodically and jumps
to a different page not staying in the loop
forever.




We term this E to be a vector over all the web
pages that accounts for each pages escape
probability (user defined parameter).
) (
) (
) ( u cE
N
v R
c u R
u
B v
v
+ =

e
PageRank Computation

- initialize vector over web pages
Loop:
- new ranks sum of normalized backlink ranks

- compute normalizing factor

- add escape term

- control parameter

While - stop when converged
S R
0
i
T
i
R A R
+1
1
1
1
+

i i
R R d
dE R R
i i
+
+ + 1 1
i i
R R
+1
o
c o >
Matrices
A is designated to be a matrix, u and v correspond to the
columns of this matrix.








Given that A is a matrix, and R be a vector over all the Web
pages, the dominant eigenvector is the one associated with
the maximal eigenvalue.
Example
A
T
=
Example (cont.)
A =
R =
Normalized =
A x = x
| A - I | x = 0
R = c A R = M R
c : eigenvalue
R : eigenvector of A
Implementation
1. URL -> id
2. Store each hyperlink in a database.
3. Sort link structure by Parent id.
4. Remove dangling links.
5. Calculate the PR giving each page an
initial value.
6. Iterate until convergence.
7. Add the dangling links.

Example
1 =
B
N
Page A Page B
Page C
2 =
A
N
1 =
C
N
Which of these three has the highest page
rank?
1 =
B
N
Page A Page B
Page C
2 =
A
N
1 =
C
N
0
1
) (
2
) (
) (
0 0
2
) (
) (
1
) (
0 0 ) (
+ + =
+ + =
+ + =
B Rank A Rank
C Rank
A Rank
B Rank
C Rank
A Rank
Example (cont.)
Re-write the system of equations as a Matrix-
Vector product.

|
|
|
|
|
|
|
.
|

\
|
|
|
|
|
|
|
|
.
|

\
|
=
|
|
|
|
|
|
|
.
|

\
|
) (
) (
) (
0 1
2
1
0 0
2
1
1 0 0
) (
) (
) (
C Rank
B Rank
A Rank
C Rank
B Rank
A Rank

The PageRank vector is simply an eigenvector
(scalar*vector = matrix*vector) of the coefficient
matrix.
Example (cont.)
1 =
B
N
Page A Page B
Page C
2 =
A
N
1 =
C
N
PageRank = 0.4
PageRank = 0.4
PageRank = 0.2
Example (cont.)
0
1
2
3
.
.
.
.
11
12
with d= 0.5
Pr(A) PR(B) PR(C)
A B
C
Example (cont.)
Other Applications
Help user decide if a site is trustworthy.
Estimate web traffic.
Spam detection and prevention.
Predict citation counts.

Issues
Users are not random walkers.
Starting point distribution (actual usage
data as starting vector).
Bias towards main pages.
Linkage spam.
No query specific rank.
References
Authoritative Sources in a Hyperlinked
Environment, Jon Kleinberg, Cornell
University.
The PageRank Citation Ranking:
Bringing Order to the Web, Lawrence
Page and Sergey Brin, Stanford
University.