What stands at the base of the most competitive search engine

Jul 17, 2008 08:44 GMT  ·  By

Many wonder what technologies and methods Google employs in order to return such accurate results to the queries. Amit Singhal, Google Fellow, has revealed some of the secrets that make 80% of the people from most countries choose Google to the detriment of its competitors. Information Retrieval, that mostly uses words salience to return results, is the lead technology that Google has been relying on for over 10 years. The guiding lines for the developers are related to the understanding of three pillars that support the system: the pages, the queries and the users.

The crawl and indexing systems, that are continuously evolving, allow developers to have at hand a "large and very fresh index." This, along with the association of important concepts to a webpage, although sometimes the query is not very relevant to what people are actually searching for, enables the efficient return of results. As the Google affiliate says, even if users search for a specific location, translated into their mother language, the results will lead to the official page.

When it comes to understanding what users want - not what users type in the search tab -, Google's developers use the "best-in-class spelling suggestion system, an advanced synonyms system, and a very strong concept analysis system." The advancement of concept identification was the milestone in Google's ranking system. Allowing to extrapolate what people are searching for, better results are impossible not to be displayed. For example, if someone types in "can bees predict rain?", the indexing system will lead users to pages such as "can animals and animal behaviors predict the weather," leading them to further analysis.

Localization is one of the impacting features used for returning the most accurate results. Google has to understand, for instance, what different nations understand of "football," because this sport has different connotations in the U.S. and the U.K., or in other countries. Personalization, meaning the ability to create a profile of the user according to his previous queries, offers some of the best results, although is it disputed. Google assures users that, by staying connected to their Gmail account, they will receive better results to their searches.

Finally, Singhal gives an actual example of one of the latest technologies used in the optimization of the search engine. Cross Language Information Retrieval employs Google's translation technologies to suggest a correspondent in English for some of the queries. The English version is offered because many of the official pages, that usually give most of the information people are searching for, are written in this language.