TrustRank Algorithm


	TrustRank Algorithm

Additional Articles:

What would happen if Google took back all of the pagerank on the web today and gave it to a small number of related websites that it trusts? These related websites could give their pagerank to their related sites and so on, much like pagerank is distributed today.

The situation above describes the TrustRank algorithm, which was proposed by three scientists from Stanford University. Instead of the democracy that prevails on the Internet today where each page has one pagerank vote, the new algorithm stands for focusing the voting ability with a small group of websites that is trusted by Google, and is allows to distribute it as it wishes. Goodbye democracy ? hello dictatorship.

TrusRank Algorithm Premise

The main premise of the TrustRank algorithm is that ?good? pages will usually link to other ?good? pages, unless they have been deceived. ?Bad? pages can definitely link to ?good? pages as an attempt to look ?good?. Therefore, the basic assumption is that the TrustRank will mostly transfer to good pages ? as good and bad pages vote for them.

The second premise is that pages that contain many outbound links pay less attention to the sites they link to. Therefore, similarly to the pagerank, a page's vote splits between all of its outbound links according to the new algorithm.

The third assumption is that the farther you are from the initial ?safe? sites set, you are more likely to encounter pages that are less ?trustworthy?. Therefore, similarly to the pagerank algorithm, the new algorithm also has a damping element that weakens every vote as it get farther from the initial ?safe? site.

An additional assumption that is presented in the article is that it's enough to use no more than 200 sites in the initial group in order to achieve good results, which means that there will be no need or only minimal need for human intervention in the process, except for maybe the initiation of the process.

Total pagerank improvement

Despite all of its complex mathematics in articles regarding the TrustRank algorithm, it is actually an improvement of the pagerank algorithm. This is probably the reason why Google was so enthusiastic about it: you don't need a new system in order to implement it! All you need to do is make small alterations to the existing system that already functions very well.

Although this time it looks like Google will not repeat the mistake it did with the pagerank and will not expose it to the public on their toolbar, in order to avoid a ?TrustRank sale? situation, where ?important? sites sell their outbound links.

Differences between TrustRank and PageRank

The main difference between the two algorithms is that the pagerank algorithm needs to be executed numerous times until relatively fixed values are received. However, the TrustRank algorithm needs to be executed only a few times in order to not allow the ?confidence votes? to get too far from the initial websites. How many times is the algorithm executed? That is not clear? but it is pretty clear that if you execute the calculations too many times, you will receive a very similar output to the current pagerank results.

Does Google use the algorithm?

It's hard to tell. In any case, it is important to remember that the TrustRank algorithm will be one of a hundred other factors in determining webpages' placement in search results. However, it might gradually replace the old pagerank algorithm.