CHECKING RELEVANCE BETWEEN KEY WORDS AND WEBSITE CONTENT Russian patent published in 2009 - IPC G06F17/30 

FIELD: physics; computer engineering.

SUBSTANCE: invention relates to data analysis and particularly to checking relevancy between terms and website content. Site content is found from the URL statement. Extended term(s), which is (are) semantically and/or contextually associated with term(s) of the statement, is (are) calculated. Measurements of content similarities and extended similarities from corresponding combinations of the term(s) of the statement, content of the site and extended terms are calculated. Measurements of category similarities between the extended terms and site content are determined, taking into account the trained similarity classifier. The trained similarity classifier is trained from the analysed site content associated with catalogue data. A certainty value is calculated, which provides an objective measure of relevancy between the term(s) of the statement and site content, from measured content similarities, extended similarities and category similarities, which evaluate several similarity evaluations taking into account the trained model of the relevancy classifier.

EFFECT: invention enables finding key words, more relevant to website content.

41 cl, 4 dwg, 1 tbl