Method for thematic classification of documents, thematic classification module and search engine incorporating such a module

Abstract From the Canadian patent

The invention concerns a method for thematic classification of documents in particular for constituting or updating thematic databases (42) for a search engine comprising steps which consist in: selecting documents representing each theme; identifying in the selected documents elements characteristic of each theme; assigning to each identified element a coefficient (R) representing the relevance of said element relative to the corresponding theme; and for each document (50) to be classified identifying said elements characteristic of each theme it contains and for each theme which corresponds to them computing from the coefficient assigned to said elements the value of a characteristic …