摘要 |
FIELD: information technology.SUBSTANCE: invention relates to systems and methods for elimination of shingles from parts of a message, which are met only in messages which do not contain spam, when filtering spam. System for eliminating shingles met only in messages which do not contain spam comprises: a) text processing means intended for: receiving a message, at least one part of text which is insignificant, wherein insignificant is part of message body, which has no value when determining spam and contains a word, symbols, based on which is separated at least an mail address, a telephone, postscriptum, auto-signature, and which is met in messages not containing spam, search in said message of said parts of text, which coincide with known parts of text from data base of samples of text, reducing text of said message by deletion of said message found parts of text, which coincide with known parts of text from data base of samples of text, sending reduced text of said message to shingles processing means; b) a data base of samples of text, intended for storage of known parts of text message, met only in messages which do not contain spam and characterise insignificant parts of message; c) shingles processing means, designed to: calculate a set of shingles based on reduced text of said message, compare calculated set of shingles with known singles from database of shingles, reducing calculated set of shingles by excluding shingles, which coincide with known shingles from database of shingles; d) database of shingles intended for storage of known shingles met only in messages which do not contain spam.EFFECT: technical result of present invention consists in reduction of size of messages when filtering spam.13 cl, 4 dwg, 2 tbl |