But if you were looking for cat food information, it would be the other way around. Relevance Google has several ways to determine the relevance of a page. At its simplest, it checks that the pages contain the same keywords as the search query. “ The simplest signal indicating the presence of relevant information in a Web page is the presence of keywords appearing in your search query.” But relevance goes much further than looking for the presence of keywords. Google also uses data on the interaction of its users with its search results. In other words, do users find the page useful? “ Beyond simply identifying common keywords, we use aggregate and anonymous interaction data to assess the relevance of search results to queries.

Google And Other Search

That’s a big part of why all the top results for “apple” are about the company, not the fruit. By analyzing the interaction data, he determind that the majority of people typing in these searches want to learn more about the brand than the fruit. But these famous interaction data are not the only elements on which Google relies. Society has investe in many different technologies to help it better understand the connections between entities like people, places and things. The Knowldge Graph is one such technology, which is no more and no less than a huge database of entities and the relationships they have between them. Apple (the fruit in English) and Apple (the company) are entities known to the Knowlee Graph.

Engines Are Not Based

Google uses the links between entities to better understand the relevance of a page. A result that matches “apple” and also talks about oranges and bananas clearly refers to the fruit. But one that also talks about iPhone, iPad and iOS certainly talks about the company. It is partly thanks to this Knowlege Graph that Google can do more than just match keywords. Sometimes you may see search results that don’t even contain the keywords that seem important in the query. For example, if we look at some first page results for “apple paper app”, the word “apple” is not in the content. Google knows this is a relevant result largely because the page mentions entities like iPhone and iPad which are obviously very close to Apple in the Knowlege Graph.

Leave a comment

Your email address will not be published.