Magazine article Information Today

Northern Light Adds Link Popularity to Its Relevance Ranking Factors List

Magazine article Information Today

Northern Light Adds Link Popularity to Its Relevance Ranking Factors List

Article excerpt

Quality is one of the most elusive elements that search engines must deliver. Web search engines in particular seek new methods for separating the best documents from the so-so and the misleading ones. Last week, Northern Light Technology, Inc. announced that it was adding link popularity to the list of factors it considers in producing a ranked list of results. Much like Google or Clever, Northern Light will note the links going to a site as an indicator of research value. This is similar to using a citation search as a method for determining which authors are most influential in a field.

Marc Krellenstein, the chief technology officer for Northern Light, was quick to point out that, unlike several search engines that appear to rely exclusively on link popularity to produce their top ranked hits, Northern Light is simply adding this factor as an additional indicator of value. Northern Light also uses the following measures (taken from the company's press release):

* Statistical measures such as query term frequency, inverse document, frequency of the term, and length of the document

* Document date

* Word context information such as whether or not the word or phrase queried occurs in the document title

* Document classification based on Northern Light's patented technology for automatically classifying the Web and organizing results into Custom Search Folders

* Natural-language analysis of the query, which analyzes the specific syntax and semantics of the user's query

Search engines add a weight for each of these sorts of measures then calculating the relevance ranking of each document. Documents that have more occurrences of the query terms, and more terms appearing close together or appearing in the title, as well as a more recent date, are given a higher ranking. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.