December 03, 2014

SEO and PageRank Algorithms Research Paper

search engine optimisation, seo tips, seo optimisation, research paper search engine, seo basics, seo tools,pagerank google, PageRank Algorithms Research Paper 

SEO

Search Engine Optimization


Search Engine Optimization and PageRank Algorithms


Nilkanth Satish Shet Shirodkar – 14103  M.Tech Goa University


Abstract: World Wide Web is a huge repository of information resources that include text, audio, video etc. As the amount of information available on web is increasing it is difficult to acquire genuine information on web. Therefore users today mainly depend upon various search engines for finding suitable answers for their queries. Search engines may return millions of pages in response to a query. It is not possible for a user to preview all the returned result set. So search engine make use of ranking algorithm to display the resultant pages in a ranked order using different page ranking algorithms. SEO is the process of getting traffic from search engines such as Google, Yahoo and Bing. local listings are shown and ranked based on what the search engine considers most relevant to users.In this paper, we will consider most popular Link based ranking algorithms namely PageRank algorithm. Relative strengths and limitations of these algorithms are explored to find out further scope of research.

Keywords: Search Engine, Worldwide web, PageRank, Ranking algorithm, Search Engine Optimization

                I.INTRODUCTION

World Wide Web is a vast resource of hyperlinked and heterogeneous information including text, audio, video and metadata. It is estimated that WWW is doubling in size every six to ten months. Due to the rapid growth of information resources on World Wide Web it is difficult to manage the information on the web. Therefore it has become necessary for the users to use efficient information retrieval techniques to find and order the desired information. Search engines play an important role in searching web pages. The search engine gathers, analyzes, organizes the data from the internet and offers an interface to retrieve the network resources. Search engines are “programs” that search documents for specified keywords and returns a list of the documents where the keywords were found. The search results are generally presented in a line of results often referred to as search engine results pages (SERPs).The major components of a Search Engine are the Crawler, Indexer, Query Processor. A crawler or spider is a program that traverses the web by following hyperlinks and storing downloaded pages in a large database. The crawler starts with a seed URL and collects documents by recursively
fetching links and storing the extracted URL‟s into a local repository. Indexer extracts the terms from each web page and records the URL where each word has occurred.Query Processor is responsible for receiving and filling search requests from the user.When a user fires a query, query engine searches the web page in the index created by the indexer and returns a list of URL‟s of the web pages that match with the user query.
In general Query Engine may return several hundreds and thousands of URL in response to a user query which includes a mixture of relevant and irrelevant information. Since no user can read all web pages returned in response to the user query, Page Ranking mechanisms are used by most search engines for putting the important pages on top leaving less important in the bottom of the result list. Popular Page Ranking algorithms used are Page Rank alogorithm, Hypertext Induced Topic Search (HITS), Weighted Page Rank algorithm, Page Content Rank etc.
           

II.RANKING ALGORITHMS
Web-page ranking  is an optimization technique used by search engines for ranking hundreds and thousands of web pages in a relative order of importance. To rank a web page different criteria are used by ranking algorithms. For example some algorithms consider the link structure of the web page while others look for the page content to rank the web page. Broadly Page Ranking algorithms can be classified into two groups Content-based Page Ranking and Connectivity-based Page Ranking. 

Content-based Page Ranking: In this type of ranking the pages are ranked based on their textual. Factors that
influence the rank of a page are :
· Number of matched terms with the query string
· Frequency of terms i.e  the number of times the search string appears in the page. The more time the string appears, the better is the page ranking
· Location of terms i.e query string could be found in the title of a page or in the leading paragraphs of a page or even near the head of a page.


Connectivity-based Page Ranking (Link based): This type of ranking work on the basis of link analysis technique. They view the web as a directed graph where the web pages form the nodes and the hyperlinks between the web pages form the directed edges between these nodes. There are two famous link analysis methods:
· PageRank Algorithm
· HITS Algorithm and

III.   PAGERANK

The  PageRank  algorithm  was  developed  at  Stanford University  by  Larry  Page  and Sergey  Brin  in  1996. PageRank algorithm , named after Larry Page and used by the Google Internet search engine uses the link structure of  the  web  to  determine  the  importance  of  the  web  page. Here a page obtains a higher rank if sum of its back-links  is high.  This algorithm is based on random surfer model. The  random  surfer  model  assumes  that  a  user  randomly keeps  on  clicking  the  links  on  a  page  and  if  he  get  bored of a page then switches to another page randomly. Thus, a user under this model shows no bias towards any page  or  link.  PageRank(PR)  is  the  probability  of  a  page  being visited by such user under this model.  For each web  page, Page Rank value is  pre-computed . For this over 25 billion web pages on the WWW are considered to assign a  rank value. A Simplified version of PageRank is defined in Equation 

 PR(A) = (1-d) + d (PR(T1)/C(T1) + ... + PR(Tn)/C(Tn)) .

Where: 
PR(A) = PageRank of page A 

T1….Tn= All pages that link to page A 

PR(Ti) = Page rank of page Ti

Q(Ti) = the number of pages to which Ti links to 

d = damping factor which can be set between 0 and 1 

PR(Ti)/Q(Ti) = PageRank of Ti distributing to all pages that Ti links to. 

(1-d) = To make up for some pages that do not have any 
out-links to avoid losing some page ranks

Damping  factor:  The  PageRank  theory  holds  that any imaginary  surfer  who  is  randomly  clicking  on  links  will eventually stop clicking. The probability, at any step,that the person will continue is called a damping factor  d. The damping  factor can be  set to any  value  such that 0<d<1, nominally  it  is  set  around  0.85.  The damping  factor  is subtracted  from  1  and  this  term  is  then  added  to  the product of the damping factor and the sum of the incoming PageRank scores.

A. Implementation of Page Rank Algorithm

The following steps explain the method for implementing 

Page Rank Algorithm.
Step  1:  Initialize  the  rank  value  of  each  page  by  1/n. 
Where  n  is  total  no. of  pages to be ranked.  Suppose we represent these n pages by an Array of n elements. Then
A[i] = 1/n where 0≤ i< n

Step  2:  Take  some  value  of  damping  factor  such  that  0<d<1. e.g 0.15, 0.85 etc. 

Step 3: Repeat for each node i such that 0≤ i< n. Let PR be an Array of n element which represent PageRank for each web page.
PR[i] <= 1-d
For all pages Q such that Q Links to PR[i] do
PR[i] <= PR[i] + d * A[Q]/Qn   
where Qn = no. of outgoing edges of Q

Step 4: Update the values of A
A[i]= PR[i]  for 0≤ i< n
Repeat  from  step  3  until  the  rank  value  converges  i.e. values of two consecutive iterations match.


B. Advantages of PageRank
The strengths of PageRank algorithm are as follows: 
·        Less  Query  time:  PageRank  has  a  clear  advantage over  the  HITS  algorithm,  as  PageRank  compute ranking at crawling time so response to user query is quick.
·        Less susceptibility to localized links:  Furthermore, as PageRank  is  generated  using  the  entire  Web  graph, rather  than  a  small  subset,  it  is  less  susceptible  to localized link.
·        More Efficient :  In contrast, PageRank computes a single  measure  of  quality  for  a  page  at  crawl  time. This  measure  is  then  combined  with  a  traditional information  retrieval  score  at  query  time.  Compared with  HITS,  this  has  the  advantage  of  much  greater efficiency.
·        Feasibility:  As  compared  to  Hits  algorithm  the PageRank  algorithm  is  more  feasible  in  today’s  scenario since it performs computations at crawl time rather than query time.

C. Disadvantages of PageRank

The  following  are  the  problems  or  disadvantages  of PageRank:
·        Less  Relevant  to  user  Query:  PageRank  score  of  apage ignores whether or not the page is relevant to the query at hand.
·        Rank Sinks: The Rank sinks problem occurs when in a network pages get in infinite link cycles
·        Spider Traps: Another problem in PageRank is Spider Traps. A group of pages is a spider trap if there are no links from within the group to outside the group.
·        Dangling Links: This occurs when a page contains a link such that the hypertext points to a page with no outgoing  links.  Such  a  link  is  known  as  Dangling Link.
·        Dead  Ends:  Dead  Ends  are  simply  pages  with  no outgoing  links.  PageRank  doesn't handle  pages  with no  outedges  very  well,  because  they  decrease  the PageRank overall.
·        Circular  References:  If  you  have  circle  references  in your  website,  then  it  will  reduce  your  front  page‟s PageRank.
·        In Internet,  available data is huge and the algorithm is not fast enough.

HITS Algorithm 

HITS  algorithm ranks the  web page by processing in links and out links of the web pages. In this algorithm a web page is  named  as  authority  if  the  web  page  is  pointed  by  many hyper  links  and  a  web  page  is  named  as  HUB  if  the  page point  to  various  hyperlinks. 

HITS  is  technically,  a  link  based  algorithm.  In HITS algorithm, ranking of the web page is decided by analyzing their textual contents against a given query. After collection of  the  web  pages,  the  HITS  algorithm  concentrates  on  the structure of the web only, neglecting their textual contents. 

Original  HITS  algorithm  has  some  problems  which  are given below. 
(i) High rank value is given to some popular website that is not highly relevant to the given query. 
(ii)  Drift  of  the  topic  occurs  when  the  hub  has  multiple topics as equivalent weights are given to all of the outlinks of  a  hub  page.  Figure  5  shows  an  Illustration  of  HITS process.

To minimize the problem of the original HITS algorithm, a clever  algorithm  is  proposed  by  reference.  Clever algorithm  is  the  modification  of  standard  original  HITS algorithm.This algorithm provides a weight value to every link depending on the terms of queries and endpoints of the link. An anchor tag is combined to decide the weights to the link  and  a  large  hub  is  broken  down  into  smaller  parts  so that every hub page is concentrated only on one topic. Another  limitation  of  standard  HITS  algorithm  is  that  it assumes equal weights to all the links pointing to a webpage and it fails to identify the facts that some links may be more important  than  the  other.  To  resolve  this  problem,  a probabilistic  analogue  of  the  HITS  (PHITS)  algorithm  is proposed by reference. A probabilistic explanation of relationship  of  term  document  is  provided  by  PHITS.  It  is able  to  identify  authoritative  document  as  claimed  by  the author.  PHITS  gives  better  results  as  compared  to  original HITS  algorithm.   Other  difference  between  PHITS  and standard HITS is that PHITS can estimate the probabilities of authorities compared to standard HITS algorithm, which can provide only the scalar magnitude of authority.

Weighted Page Rank Algorithm 

Weighted Page Rank Algorithm is proposed by Wenpu Xing  and  Ali  Ghorbani.  Weighted  page  rank  algorithm (WPR)  is  the  modification  of  the  original  page  rank algorithm.  WPR  decides  the  rank  score  based  on  the popularity  of  the  pages  by  taking  into  consideration  the importance  of both the  in-links  and out-links  of  the  pages. This  algorithm  provides  high  value  of  rank  to  the  more popular pages and does not equally divide the rank of a page among   its  out-link  pages.  Every  out-link  page  is  given  a rank  value  based  on  its  popularity.  Popularity  of  a  page  is decided  by  observing  its  number  of  in  links  and  out  links. 


Weighted Links Rank Algorithm 

A modification of the standard page rank algorithm is given by Ricardo Baeza-Yates and Emilio Davis named as weighted  links  rank  (WLRank).  This  algorithm  provides weight value to the link based on three parameters i.e. length of  the  anchor  text,  tag  in  which  the  link  is  contained  and relative  position  in  the  page.  Simulation  results  show  that the results of the search engine are improved using weighted links.  The  length  of  anchor  text  seems  to  be  the  best attributes in this algorithm.  Relative position, which reveal that physical position does not always in synchronism with logical position is not so result oriented. Future work in this algorithm includes, tuning of the weight factor of every term for further evolution.

EigenRumor Algorithm 

As  the  number  of  blogging  sites  is  increasing  day  by  day, there  is  a  challenge  for  service  provider  to  provide  good blogs to the users. Page rank and HITS are very promising in providing the rank value to the blogs but some limitations arise, if  these  two  algorithms  are  applied  directly  to  the blogs The rank scores of blog entries as decided by the page rank  algorithm  is  often  very  low  so  it  cannot  allow  blog entries  to  be  provided  by  rank  score  according  to  their importance.   To  resolve  these  limitations,  a  EigenRumor algorithm  is  proposed  for  ranking  the  blogs.  This algorithm provides a rank score to every blog by weighting the  scores  of  the  hub  and  authority  of  the  bloggers depending on the calculation of eigen vector.

Time Rank Algorithm 

An algorithm named as TimeRank,  for improving  the rank score by using the visit time of the web page is proposed by H Jiang  et al. Authors have  measured the visit time  of the  page  after  applying  original  and  improved  methods  of web  page  rank  algorithm  to  know  about  the  degree  of importance  to  the  users.  This  algorithm  utilizes  the  time factor to increase the accuracy of the web page ranking. Due to the methodology used in this algorithm, it can be assumed to be a combination of content and link structure. The results of this algorithm are very satisfactory and in agreement with the applied theory for developing the algorithm.
Search Engine Optimization
SEO stands for “search engine optimization.” It is the process of getting traffic from the “organic,” “editorial” or “natural” listings on search engines. All major search engines such as Google, Yahoo and Bing have such results, where web pages and other content such as videos or local listings are shown and ranked based on what the search engine considers most relevant to users.

Ways to improve your Web pages for SEO
·        URLs
Create clean, focused and optimized URLs. Include target keywords, avoid long or automated dynamic URLs.
·        <title> tag
Optimal length for search engines is roughly 70 characters, place important keywords towards the front of the title tag, include your brand name at the end of a title tag, and make sure to consider readability and emotional impact.
·        <meta>
Optimal length for search engines is roughly 155 characters, make it compelling and unique to each page.
·        <h1>
Make sure your primary keyword or phrase is enclosed in an H1 tag, make sure it’s a strong headline that reads naturally, place it near the top of your page after your opening tag.
·        Keyword ratio
This is the percentage of times a keyword is used in a body of text, the usual amount if 2-3%.
·        SiteMap
Make sure to include a detailed and accurate XML sitemap in order to show search engine spiders and your visitors where to find information on your site.


SEO Tools:
·        Raven Tools - All-in-one Internet marketing platform for SEO, Social Media, PPC and Content. Research, manage, monitor and report on every aspect of your campaign. www.raventools.com/SEO-Software
·        RIO SEO - SEO platform to deliver global search success across Organic, Local Search, Mobile and Social Media for Top Brands & Agencies. www.rioseo.com
·        Google Analytics - lets you measure your advertising ROI as well as track your Flash, video, and social networking sites and applications. www.google.com/analytics
·        SEO analytics software - SEOmoz tools helps you handle everyday SEO tasks. You can analyze keywords, research backlinks, do on-page analysis, find accessibility issues and track rankings all in one easy-to-use management platform. www.seomoz.org/tools
·        Traffic Travis - For all your SEO & PPC Management needs. Use Traffic Travis for both on and off page analysis as well as spying on your competitors. www.traffictravis.com
·        SheerSEO - SEO software. Complete automation of SEO including tools for tracking, back links building and much more.             
            
CONCLUSION

Based on the algorithm used, the ranking algorithm provides a  definite  rank  to  resultant  web  pages.  A  typical  search engine should use web page ranking techniques based on the specific needs of the users. After going through exhaustive analysis of algorithms for ranking of web pages against the various parameters such as methodology, input parameters, relevancy  of  results  and  importance  of  the  results,  it  is concluded  that  existing  techniques  have limitations 
particularly  in  terms  of  time  response,  accuracy  of  results, importance  of  the  results  and  relevancy  of  results.  An efficient web page ranking algorithm should meet out these challenges  efficiently  with  compatibility  with  global standards of web technology. 


REFERENCES

I.       Sergey Brin and  Larry Page, “The  anatomy  of  a  Large-scale Hypertextual  Web  Search Engine”, In Proceedings of the Seventh International World Wide Web Conference, 1998.

II.      Hyperlink based search algorithms-PageRank and HITS By Shatakirti

III.     Search Engine Optimization Starter Guide By Google. 

IV.     A Comparative Analysis of Web Page Ranking  Algorithms by Dilip Kumar Sharma and A. K. Sharma.


No comments:
Write comments