Friday, November 9, 2007

Biases and Restrictions for Google Search

Google uses many signals to rank search results and, in some cases, it filters some of them based on your location, device or preferences. Here are some ways to disable these filters or to create custom filters.

1. Google automatically expands some queries and also includes results that contain different morphological forms or related words. If you search for [av], Google will highlight antivirus and anti-virus. To prevent this, add + in front of the word: [+av]. This is also useful if a certain keyword is very important and must be included in each search result.

2. Google filters duplicate search results, so if you want to find a better estimation for the number of search results add &filter=0 at the end of Google's URL. This parameter disables the following two filters:
* Duplicate Content — If multiple documents contain the same information, then only the most relevant document of that set is included in your search results.

* Host Crowding — If there are many search results from the same site, Google may not show all the results from that site or may show the results lower in the ranking than they otherwise would have been.

3. Here are two useful URLs for Google's homepage:

* http://www.google.com/webhp - if Google redirects you to the mobile version, but you want the standard homepage, this URL bypasses the detection

* http://www.google.com/ncr - if Google redirects you to a localized version, based on your IP address or your browser's settings, you can go to the global version using this URL (or by clicking on "Google.com in English").

4. To introduce a local bias, add the gl parameter to Google's search results URL. For example, http://www.google.com/search?q=roma&gl=it shows the results for [Roma], but gives better rankings to pages related to Italy.

5. If you don't like Google's personalized results you can log out from your Google account, disable the Web History service or turn off the personalization bias by adding &pws=0 to Google's URL. Note that the parameter is not persistent and it only works for the current search.

6. To restrict your search to the high-quality (?) web sites included in the Open Directory Project, you can append &cat=gwd/Top to Google's URL or perform your search at Google Directory.

7. Find the right custom search engine for your query and enter your query there. A good custom search engine restricts the search space to a number of authoritative sites from a domain.

8. Use the site: operator to restrict your search to a TLD (like .edu), domain, subdomain or even a pattern. For example, to restrict your search to YouTube pages that contain videos, try: [site:youtube.com/watch/v?=].

9. Search only the web pages and sites linked from a page using Google Co-op's on-the-fly feature. Use the "try it out" option.

10. A version of Google which identifies and prioritizes "search results that are more easily usable by blind and visually impaired users" is Accessible Web Search.

Other filters are available in the advanced search page: restrict the search results to pages written in a language, from a certain country, to non-adult sites, or to pages licensed using a flavor of Creative Commons.

No comments:

Post a Comment