FILTERING UNAUTHORIZED MESSAGES IN POSTAL ELECTRONIC SERVICES

In article the solution of the task of filtering electronic mail correspondence on the basis of preliminary intellectual processing of electronic messages with use of the neural network qualifier is proposed. Processing of electronic messages includes hands-off processing of the text on the basis of linguistic approach. In operation the vectorial model of display of signs of the electronic message is considered. It is offered to use a weighing Ltcmeasure as a measure of the significance of terms. Combined approach of abbreviation of character space by calculation of the value characterizing the significance of a term for a certain class k and formation of a collocate of the message with use of indices of force of semantic (synoptic) communication between qualitative characters (terms) of phrases is also reasonable. Use of a measure of narrowness of correlation of two qualitative characters of phrases is reasonable: coefficients of association of KA and kontingention Kk. For the decision of the task of filtering unauthorized electronic messages the adaptive neural ART network by advantage of which is selected the ability to samoobuchatsya (to create images) for adaptation to the changing needs of the addressee of correspondence is the efficiency of the offered model of the electronic message integrated with method of neural network classification for intellectual filtering electronic correspondence is probed and confirmed.

1. Nikolaev, I.A. Spam: economic losses: Analytical report / [Electronic resource] (http://www.fbk.ru/news/5419/83743/).

2. Slepov, O. Content filtering / O. Slepov // JetInfo № 10 (149) / 2005 [Electronic resource] (http://www.jetinfo.ru/Sites/new/Uploads/ 2005_10.pdf).

3. Soloviev, N.A. The development of the concept of intrusion detection / ЕН. Chernoprudova, N.A. Solov'ev // Modern Information Technologies in Science, Education and Practice: Materials of the VIII All-Russia. scientificpractical. Conf., / Orenburg. state. un-t. - Orenburg: GOU, 2009. - P. 66-67. - ISBN 978-5-7410-0975-8

4. Chernoprudova, E.N. Neural network model of intellectual filtration of unauthorized mailings / E. Chernoprudova // Information Materials of the IX All-Russian Scientific and Technical Conference - Orenburg: OSU, 2010, p. 44- 47.

5. Chernoprudova, E.N. Intellectual filtering of unauthorized mailings based on a neural network / Е.Н. Chernoprudova, N.A. Solovyov // Academic Journal of Intellect. Innovation. Investments". Specialist. issue, 2011. p.106-107

6. McCallum, A. A comparison of Event Models for Naïve Bayes Classifiection / A. McCallum, K. Nigam; // In AAAI-98 Workshop on Learning for Text Categorization. – 1998 – 8 c.

7. Fuernkranz, J. A study using n-gram features for Text Categoriza-tion / J. Fuernkranz // Tech report OEFAI-TR-98-30 – 1998.

8. Dasigi, V. Neural Net Learning Issues in Classification of Free Text Documents / V. Dasigi, R. Manu // AAAI spring symposium on Machine Learning in Information Access – 1996.

9. Li, Y. Classification of Text Documents /Y.H. Li, A.K. Jain //The Computer Journal, Vol. 41, No. 8, 1998

10. Mingyong, L. An improvement of TFIDF weighting in text categorization / L. Mingyong, Y. Jiangang [Электронный ресурс] http:// www.ipcsit.com/vol47/009-ICCTS2012-T049.pdf)

11. Cover, T. Elements of Information theory / T. Cover, J. Thomas [Электронный ресурс] (https:// web.cse.msu.edu/cse842/Papers/CoverThomas-Ch2.pdf)

12. Kondratiev, M.E. Two-level hierarchical clustering of the news flow in ROMIP 2006 / M.E. Kondratiev // Russian seminar on the evaluation of methods of information retrieval. Proceedings of the fourth Russian seminar ROMIP'-2006. St. Petersburg: NU CSI, 2006, 274 p. 126-138 [Electronic resource] (http://romip.narod.ru/ romip2006 / index.html)

13. Hotho, A. Ontology-based Text Clustering / A. Hotho, S. Staab, A. Maedche [Электронный ресурс] (http://www.cs.cmu.edu/mccallum/textbeyond/papers/hotho.pdf)

14. Lan, M. Supervised and Traditional Term Weighting Methods for Automatic Text Categorization / M. Lan, C. L. Tan, Senior Member [Электронный ресурс] (https://www-old.comp.nus.edu.sg/~tancl/pub lications/j2009/PAMI2007- v3.pdf)

15. Manning, K.D. Introduction to Information Search / K.D. Manning, P. Raghavan, H. Schutze, Williams. Moscow: 2011, 528 p.

16. Yagunova, E.V. From collocations to constructions / E.V. Yagunova, LM Pivovarova // Russian language: constructional and lexical-semantic approaches - St. Petersburg, 2011. - 43с.

17. Khokhlova, M.V. Experimental verification of methods for isolating collocations. Khokhlova [Electronic resource] (http://www.helsinki.fi/ slavicahelsingiensia / preview / sh34 / pdf / 21.pdf).

18. Minashkin, V.G. Theory of statistics / V.G. Minashkin, RA Shmoilova, N.A. Sadovnikova, L.G. Mosesikina, E.S. Rybakova, Moscow: ed. the EAOI Center. 2008, 296 pp.

19. Chernoprudova, E. N. Protection of postal services from unauthorized mailings based on content filtering of electronic messages: dis. ... cand. those. Sciences: 05.13.19 / Chernoprudova Elena Nikolaevna. - Ufa, 2013. - 131 p.

20. Khaikin, S. Neural networks: full course / C. Khaikin - M: Williams, 2006, - 1104 p.

21. Certificate of state registration of the computer program No. 2013617655 "A program for spam filtering based on a neural network" / E. N. Chernoprudova, N. A. Soloviev, V. A. Puchkov. Application No. 2013617655. Date of receipt May 21, 2013 Registered in the Register of Computer Programs August 21, 2013

22. Valeev, S.S. Multilevel system of spam filtration based on artificial intelligence technologies / SS. Valeev, A.P. Nikitin // Bulletin of the USATU, 2008, vol. 11, No. 1 (28). Pp. 215-219.