Talk:PageRank

Removed

Google uses an automated web spider called Googlebot to actually count links and gather other information on web pages.

It is by no means clear that the counting can be said to be done by Googlebot, and it is not intuitively a spidering operation, more likely a feature of the database to which the spidering software stores its flies. Therefore this needs a citation to be in the article. Clearly what parts of the Google infrastructure are called "Googlebot" is up to Google, however if it extends too far, the description needs to be changed. All the best: Rich Farmbrough, 13:25, 11 August 2014 (UTC).

External links modified

Hello fellow Wikipedians,

I have just added archive links to one external link on PageRank. Please take a moment to review my edit. You may add {{cbignore}} after the link to keep me from modifying it, if I keep adding bad data, but formatting bugs should be reported instead. Alternatively, you can add {{nobots|deny=InternetArchiveBot}} to keep me off the page altogether, but should be used as a last resort. I made the following changes:

• Attempted to fix sourcing for http://www.google.com/competition/howgooglesearchworks.html

When you have finished reviewing my changes, please set the checked parameter below to true or failed to let others know (documentation at {{Sourcecheck}}).

This message was posted before February 2018. After February 2018, "External links modified" talk page sections are no longer generated or monitored by InternetArchiveBot. No special action is required regarding these talk page notices, other than regular verification using the archive tool instructions below. Editors have permission to delete these "External links modified" talk page sections if they want to de-clutter talk pages, but see the RfC before doing mass systematic removals. This message is updated dynamically through the template {{source check}} (last update: 5 June 2024).

• If you have discovered URLs which were erroneously considered dead by the bot, you can report them with this tool.
• If you found an error with any archives or the URLs themselves, you can fix them with this tool.

Cheers.—^{cyberbot II}_{Talk to my owner:Online} 05:05, 31 March 2016 (UTC)

• What does the link: Our products and services by Google have to do with PageRank? I followed the link but could not find PageRank. If it is somewhere at that URL, then the link should be refined. 5.34.22.174 (talk) 04:41, 5 January 2018 (UTC)

Minor error in "Damping Factor"

In "Damping Factor", after the two formulas, it states, "The difference between them is that the PageRank values in the first formula sum to one, while in the second formula each PageRank is multiplied by N and the sum becomes N." However (unless I'm blind), from the first equation to the second, only the first part has been multiplied by N. (Otherwise you'd have 1 - d + Nd(stuff).) With the second equation given as-is, the sum of page ranks is rather trickier than "N". Given that the equation is already acknowledged to be wrong, it's probably not urgent, but hey. — Preceding unsigned comment added by 98.102.161.228 (talk) 17:37, 22 August 2016 (UTC) The parameter of this code is good to be used in many problems..hoursguru.com

Damping factor for biological data

The article mentions 0.31 as the optimal value for d, however nowhere in the cited paper can I directly find 0.31. I can find a reference to an epsilon of 0.3 in Appendix 1.1, but I am not yet convinced this is equivalent to the damping factor d. Can someone clarify? --José Devezas (talk) 08:19, 19 December 2018 (UTC)

I also wonder whether the statement even makes sense. How can there be a single damping factor for all "biological data"? What application are we talking about? What is being modelled? I will remove the statement. If someone feels it adds value or explanatory power to the article, perhaps they can clarify it. GBMorris (talk) 13:25, 9 March 2020 (UTC)

I agree with the points raised by both of you and consider the removal of the statement a good decision. BernardoSulzbach (talk) 20:50, 9 March 2020 (UTC)

Examples are Wrong

• It's not clear what does the value 0.5 represents
• The sum(i, M_i,j) is never 1
• I think columns/rows have been messed up. Comments mention that M_i,j represents link from j to i, but if you run a clear example, e.g, the following:

everything links to A which means A is an important page, and A links to C, thus C is important as well.

   M = np.array([
       [1, 1, 1],  # * -> A
       [0, 0, 0], 
       [1, 0, 0]   # A -> C
   ])
  print(pagerank(M, 0.001, 0.85)) 
  array([[0.61536926],  
         [0.28131799],
         [0.10331275]])

B should be last but its second. Am I missing something?

I'm not sure which version of the algorithm this edit was referring to, but I've been working with it and it works; in this case, [1, 1, 1] should be impossible because A cannot link to itself. A links to C would be [0, 0, 1] on the A row if [A, B, C], and A links to both B and C would be [0, 1, 1]. I'm also not sure what the call to `pagerank(..., 0.001, ...)` is meant to do, since `0.001` doesn't make sense as the number of iterations (I'm guessing it would iterate once).

--D.g.lab. (talk) 16:31, 10 April 2021 (UTC)

I'm also not sure what was meant by the sum of (i, M_i, j) is 1 since half those variable names don't exist in the code. It could be said that the sum of every column in the matrix is 1?

The sum of the result vector, called v in the code, should also be approximately 1, I believe, since it contains percentages.

--D.g.lab. (talk) 16:39, 10 April 2021 (UTC)

Variables lack clear names

It's already difficult enough to understand as it is, why not clarify the variable names and add comments, as in any good code? For example:

   def pagerank(matrix, num_iterations: int = 100, damping: float = 0.85):
       size = matrix.shape[1]
       res_vector = [1/size] * size
       M_hat = (damping * matrix + (1 - damping) / size)
       for i in range(num_iterations):
           res_vector = M_hat @ res_vector # matrix multiplication
       return res_vector

Also, notice the lines in the article:

   v = np.random.rand(N, 1)
   v = v / np.linalg.norm(v, 1)

these insert random numbers in the multiplication vector. However, as far as I understand, this vector is supposed to start initially as `1/number_of_pages` (1/size) for each of its elements, which could be re-written

   v = [1/N] * N

or, using the new varible names,

   res_vector = [1/size] * size

According to this video on YouTube: Linear Algebra – Introduction to PageRank (4:21 timestamp).

I'm unsure why this was like this, so I'm not editing the page; however, if someone can confirm, the initial method seems "way too complicated" and there's no explanation for it. --D.g.lab. (talk) 16:31, 10 April 2021 (UTC)

site:example.com

😐site:example.com 2A00:F41:4828:36B6:0:4B:5D7E:4801 (talk) 23:52, 24 February 2022 (UTC)

reference 25 is old

the article states:'PageRank continues to be the basis of googles....'[25] reference 25 is over 10y old so it does not prove anything. 213.247.64.239 (talk) 14:32, 18 January 2023 (UTC)

@213.247.64.239

actual text from the article:

'While just one of many factors that determine the ranking of Google search results, PageRank continues to provide the basis for all of Google's web-search tools.' 213.247.64.239 (talk) 14:33, 18 January 2023 (UTC)

Google doesn't disclose the details of its ranking systems, but they have affirmed that PageRank remains one of the many signals they use. While there might be a more recent reference, it wouldn't be very precise. They don't disclose those kinds of details. While this article could use some clean up, it's not inaccurate in saying that Google still uses the algorithm. Michael Martinez (talk) 15:01, 18 January 2023 (UTC)

Google recently added new documentation that confirms PageRank is still used in their ranking systems although it has evolved since the original paper (in unspecified ways).Michael Martinez (talk) 13:01, 25 April 2023 (UTC)

"Supplemental Result" listed at Redirects for discussion

The redirect Supplemental Result has been listed at redirects for discussion to determine whether its use and function meets the redirect guidelines. Readers of this page are welcome to comment on this redirect at Wikipedia:Redirects for discussion/Log/2024 April 25 § Supplemental Result until a consensus is reached. Utopes _{(talk / cont)} 05:12, 25 April 2024 (UTC)

Merge PageRank algorithm in biochemistry

Most of the content on PageRank algorithm in biochemistry is redundant to material here, including a confusing and incomplete explanation of how PageRank works. It also hasn't had any meaningful updates since 2014. The small amount of material that actually pertains to PageRank's biochem applications should be merged here. Apocheir (talk) 01:56, 23 June 2024 (UTC)

I'm not sure merging the articles is the right thing to do. PageRank is based on citation analysis. To date no one has suggested merging PageRank into the Citation analysis - Wikipedia article. And I don't think that would be appropriate, either. While I agree there is some redundancy between the two PageRank articles, there are similar redundancies across Wikipedia. Cross-linking them and minimizing the duplication is the better solution in my opinion. Michael Martinez (talk) 03:08, 23 June 2024 (UTC)

I would recommend just taking this to AFD. InfiniteNexus (talk) 08:45, 26 June 2024 (UTC)