METHOD OF SEARCHING FOR SIMILAR ELECTRONIC DOCUMENTS STORED ON DATA STORAGE DEVICES Russian patent published in 2015 - IPC G06F17/30 G06F17/27 

FIELD: information technology.

SUBSTANCE: method of searching for semantically similar electronic documents stored on data storage devices includes loading two electronic documents; determining search parameters by setting rules for generating a plurality of unique words which form a plurality of weighted unique words and weighted links between said words; constructing a semantic network and searching for semantically similar documents by comparing the semantic networks; additionally setting rules for generating stylistic images of documents by determining the size of transition frequency matrices and selecting elements of the transition frequency matrices, wherein elements of the transition frequency matrices either bigrams or trigrams; generating a transition frequency matrix of documents and comparing the transition frequency matrix of documents for similarity by calculating a similarity coefficient.

EFFECT: higher accuracy of searching for similar electronic documents in an array of documents of different styles.

2 dwg