Books

S. Raudys. (B2001 ) Statistical and Neural Classifiers: An integrated approach to design. Springer. London. 312 pages.

S. Raudys. (B1984) Statistical Pattern Recognition: Small design sample problems. A monograph, (a manuscript ) Institute of Mathematics and Cybernetics, Vilnius, 480 pages, 30 copies distributed around the world.

S. Raudys. (B1976) Limitation of Sample Size in Classification Problems, Inst. of Physics and Mathematics Press, Vilnius. 186 pages.

 Book Chapters

S.Raudys and Jain K. (1991a). Small sample size problems in designing Artificial  Neural Networks. - Artificial Neural Networks and Statistical Pattern Recognition, Old and New Connections, I.K. Sethi and A.K. Jain (Eds), Elsevier Science Publishers B.V, 33-50.

S.Raudys, (1978a) Optimization of nonparametric classification algorithm. Adaptive systems and applications .  Nauka, Novosibirsk, (A.Medvedev Ed.), 57-62.

Journal Articles

S. Raudys, A. Saudargiene (2001). Tree type dependency model and sample size - dimensionality properties. IEEE Transactions on Pattern Analysis and Machine Intelligence 23(1), 233-239.

V. Diciunas, S. Raudys (2000). Generalization error of randomized linear zero empirical error classifier: Simple asymptotics for centered data case. Informatica, 11 (4), 381-396 .

S. Raudys. (2000c). Scaled rotation regularization. Pattern Recognition 33, 1989-1998 .

M. Skurichina, S. Raudys, RPW. Duin (2000). K-nearest neighbors directed noise injection in multilayer perceptron training, IEEE Trans. on Neural Networks, 11, 504-511.

S. Raudys (2000b). How good are support vector machines? Neural Networks 13, 9-11.

S. Raudys (2000a). Evolution and generalization of a single neurone. III. Primitive, regularized, standard, robust and minimax regressions. Neural Networks 13 (3/4), 507-523.

S. Raudys(1998b).Evolution and generalization of a single Neurone. II. Complexity of statistical classifiers and sample size considerations. Neural Networks, 11( 2), 297-313.

S .Raudys (1998a). Evolution and generalization of a single neurone. I. SLP as seven statistical classifiers. Neural Networks, 11( 2), 283-296.

S. Raudys, R.P.W. Duin (1998). On expected classification error of the Fisher classifier with pseudo-inverse covariance matrix. Pattern Recognition Letters, 19, 385-392.

S. Raudys (1998c). Intrinsic dimensionality and small sample size properties of classifiers. Kybernetika, 34, N 4, 461-466.

E. Cagatay Guler, B. Sankur, Y. P. Kahya, S.Raudys (1998). Visual classification of medical data using MLP mapping, Computers in Biology and Medicine, 28,  272-287.

S. Raudys. (1997a ). On  dimensionality,  sample size  and  slassification error of nonparametric linear  classification  algorithms", IEEE Transaction on Pattern Analysis and Machine Intelligence. PAMI-19 (6), 669-671.

A. Basalykas, V. Diciunas, S. Raudys (1996+1997). On expected probability of misclassification of zero empirical error classifier. Informatica, 7(2), 137-154, 8(2), 310-311.

T. Cibas, F. Fogelman, P. Gallinari and S. Raudys (1996). Variable selection with neural networks, International  J.   Neurocomputing , 12, 223-248.

S. Raudys, M. Skurikhina, T. Cibas, P. Gallinari (1995). Ridge estimates of the covariance matrix and regularization of artificial neural network classifier. Pattern Recognition and Image Processing, Int. J. of Russian Academy of Sciences, Moscow, N4, 633-650.

S.Raudys, (1993). On shape of pattern error function, initializations and intrinsic dimensionality in ANN design, Informatica, 4(3-4), 360-383.

A. Jain and S. Raudys (1992). On training sample size and complexity of artificial neural net classifier. Informatica,  3(3), 301-337.

S. Raudys and  A.K. Jain (1991b).  Small  sample  size  effects in statistical  pattern  recognition: Recommendations for  practitioneers. - IEEE Transactions on Pattern Analysis and Machine Intelligence, PAMI -13 (3), 252-264.

S.Raudys (1991b). Methods to  overcome  dimensionality  problems  in statistical pattern recognition. An invited review  paper.  Zavodskaya Laboratorya (USSR inerdisciplinary Journal), Moscow, Nauka, No 3, 45, 49-55 (in Russian).

S. Raudys (1991c).  On the efectiveness  of  Parzen  window  classifier,   Informatica, 2(3), 434-454.

S. Raudys, (1990). A state of art in statistical pattern recognition and unsolved problems,  - Statistical Problems of Control, 93, 9-23.  (in Russian).

S. Raudys (1986a ). The accuracy of model selection in data analysis, - Statistical Problems of Control, 74, 9-21 (in Russian).

S. Raudys, V. Pikelis, D. Stasaitis (1986). The effect of the number of initial and final number of features, the dependence between the features and a type of the classification rule on the accuracy of feature selection. - Statistical Problems of Control, 74, 39-47 (in Russian).

V. Pivoriunas and S. Raudys (1986). Classification objects into classes described by different sets of variables. - Statistical Problems of Control, 74, 119-131 (in Russian).

S. Raudys (1984b ). The influence of sample size on classification performance(a review),  - Statistical Problems of Control, 66, 9-24 (in Russian).

S. Raudys and V. Vaitukaitis (1984). Methods to estimate the probability of misclassification, - Statistical Problems of Control, 66, 43-65 (in Russian).

V. Pivoriunas and S. Raudys (1984d) Classification rule when mean vectors of the classes and object have unequal number of components. - Statistical Problems of Control, 66, 121-136 (in Russian).

V. Pikelis and S. Raudys (1982a).  General description of the program package SORRA -2. Input language. Usage. - Statistical Problems of Control, 58, 9-26 (in Russian).

S.Raudys and V. Pikelis (1982b ). Collective selection of the best version of a pattern recognition system. - Pattern Recognition Letters, 1(1), 7-13.

S. Raudys (1982).Solving  of data analysis problems by means of the applied program package SORRA-2.- Statistical Problems of Control, 58, 67-76 (in Russian).

S. Raudys (1981).  Influence of sample size on the accuracy of model selection in pattern recognition. - Statistical Problems of Control, 50, 9-30 (in Russian).

V. Vaitukaitis, V. Pikelis, S. Raudys (1981). Comparison of some sets of measurements in acoustic signal recognition. - Statistical problems of control, 50, 91-98 (in Russian).

S. Raudys and V. Pikelis (1980). On dimensionality, sample size, classification error and complexity of classification agorithm in pattern recognition. - IEEE Trans. on Pattern Analysis and Machine Intelligence, PAMI-2 (3), 242-252.

S. Raudys (1979a ). Classification errors when features are selected. - Statistical Problems of Control, 38, 9-25 (in Russian).

S. Raudys (1979b ). Determination of optimal dimensionality in statistical pattern classification. - Pattern Recognition, 11, 263-271.

D. Griðkevièius and S. Raudys (1979). On the expected probability of classification error of the classifier for discrete variables, - Statistical Problems of Control, 38, 95-112 (in Russian).

S. Raudys (1978). Classification and regression subroutines - Statistical Problems of Control, 27 , 23-40 (in Russian).

V. Pivoriunas and S. Raudys (1978).  On the accuracy of "leaving-one-out" estimate. - Statistical Problems of control, 27, 53-70, (in Russian).

S. Raudys (1976b ). Classification errors of the quadratic discriminant function. - Statistical Problems of Control, 14,  Vilnius, 33-48 (in Russian).

S. Raudys (1976d ). Investigation of a nonparametric classifers when sample size is limited. - Statistical Problems of Control, 14, 117-126 (in Russian).

S. Raudys (1976c ). Determination of a number of useful measurements in pattern recognition. - Statistical Problems of Control, 14, 137-150 (in Russian).

K.  Juskevicius and S. Raudys (1976). Estimation of the probability of misclassification for a piecewise decision rules. - Statistical Problems of Control, 14, 67-84 (in Russian).

A. Baublys and S. Raudys (1975).  Mathematical models for storage of statistical data. - Statistical Problems of Control, 13, 97-120 (in Russian).

S. Raudys, V. Pikelis and K. Juskevicius (1975). Experimental comparison of thirteen classification algorithms. - Statistical Problems of Control , 11, Vilnius, 53-80 (in Russian).

S. Raudys (1975). Algorithms to design pattern classification rules (a review). - Statistical Problems of Control, 11, 11-52 (in Russian).

S. Raudys and V. Pikelis (1975). Tabulation of the probability of misclassification for the linear discriminant function. - Statistical Problems of Control, 11, 81-120 (in Russian).

A. Baublys, S. Raudys (1975). A Bayes approach to estimate probability distribution function of random factors at enterprise. Upravliajuschie Systemy i Mashiny. N 2, 117-121 (in Russian).

S. Raudys (1973a ). Comparison of two ways to estimate the probabilities of classification errors. Automation and Remote Control (Avtomatika i Telemechanika, Nauka), N10, 45-68 (in Russian).

S. Raudys.(1973b ). Estimation of probability of misclassification. - Statistical Methods of Control, 5, 9-44 (in Russian).

S.  Raudys (1973c). On the problem of selecting a classification rule. - Statistical Problems of Control, 5, 49-69 (in Russian).

J. Radvilavicute and S. Raudys (1973). Investigation of classifier, designed with the help of histograms. - Statistical problems of control, 5 , 126-139 (in Russian).

S. Raudys (1972). On the amount of a priori information in designing the classification algorithm. - Proceedings of Acad. of Sciences of the USSR, Tech. Cyber., No. 4, 168-174 (in Russian).

S.  Raudys (1971). Classification of objects with dependent means. Statistical problems of Control, 1, 87-95 (in Russian).

S. Raudys (1970). On the estimation of the distribution function of the classification error. Int. J. Automatics and Computer Science, Riga, 92-95 (in Russian).

S. Raudys (1967). On determining training sample size of  a linear classifier. - Computing Systems, Issue 28, Novosibirsk, 79-87 (in Russian).

Papers in Conference Proceedings

S.Raudys (C2000c). Classifier's complexity control while training multilayer perceptrons. In: Ferri F, Pudil P (eds) Advances in Pattern Recognition, Springer Lecture notes in computer science. Vol. 1876, pp. 32-44 (Proc. SPR+SSPR'2000, August 29-September 1, 2000. Alicante, Spain, invited paper).

S.Raudys, M.Tamosiunaite (C2000). Biologically inspired architecture of feedforward networks for signal classification. In: Ferri F, Pudil P (eds) Advances in Pattern Recognition. Springer Lecture notes in computer science. Vol. 1876, pp. 727-736 (Proc. SPR+SSPR'2000, August 29-September 1, 2000. Alicante, Spain).

S.Raudys (C2000b). Prior weights in adaptive pattern classification. Proc. of IPRS 15-th International Conference on Pattern Recognition, September 3-8, 2000. Barselona, IEEE Publication, Los Alamitos, CA, USA, Vol. 2, pp. 1014-1017.

S.Raudys(C2000a). Marrying Statistics and Neural Networks. Tutorial. 15-th International Conference on Pattern Recognition, September 2, 2000. Barselona, 52 pages.

S. Raudys, A. Saudargiene (C2000). A tree-type dependence model in statistical and neural classification. Proc. of the 5th Conference Neural Networks and Soft Computing, (Zakopane, Poland, June 6-10, 2000), 273-278.

S. Raudys (C1998). Use of statistical hypothesis in neural network design, Proc. 5th International Conference on Machine Modelling. Minsk, June 1998, 51-59 (invited paper).

 S. Raudys, A. Saudargiene (C1998).  Structures of the covariance matrices in the classifier design. Advances in Pattern Recognition. Springer Lecture notes in computer science. Vol. 1451 (Proc. joint IAPR int. workshops / SSPR'98 and SPR'98, Sydney, Australia, August 11-13, 1998), 583-592.

S. Raudys (C1998e ). Choice of parameter estimation scheme to design a linear classifier. 7-th Vilnius Conference in Probability Theory, 22th European meeting of statisticians. Abstracts. TEV, Vilnius, August 1998, 387-388.

S. Raudys, S.Amari (C1998). Effect of initial values in simple perception. Proceedings 1998 IEEE World Congress on Computational Intelligence, IJCNN'98, Anchorage, Alaska, May 1998, 1530-1535.

S. Raudys (C1997b). Intrinsic dimensionality and small sample properties of statistical classifiers. Proceedings of STIPR'97 Workshop, Praha, June 1997, 31-36.

S. Raudys (C1996a). Overtraining in single-layer perceptrons, Courses and Lectures No. 382. Learning, Networks and Statistics (Proc. of Int. Conference, Udine, Italy, September 1996, invited paper), Springer, Wien - New York, 3-24.

S. Raudys (C1996b ). Linear classifiers in perceptron design, ICPR13, Proc. 13th Int. Conf. on Pattern Recognition (Vienna, Austria, Aug.25-29) Vol. 4, Track D: Parallel and Connectionist Systems, IEEE Computer Society Press, Los Alamitos, 763-767.

S. Raudys, V. Diciunas (C1996). Expected error of minimal empirical error and maximal margin classifiers, ICPR13, Proc. 13th Int. Conf. on Pattern Recognition (Vienna, Austria, Aug.25-29) Vol. 2, Track B: Pattern Recognition and Signal Analysis, IEEE Computer Society Press, Los Alamitos, 875-879.

S. Raudys, T.Cibas (C1996). Regularization by early stopping in single layer perceptron training, Proceedings of ICANN'96, Bohum, Germany, Eds. C.von der Malsburg, W.von Seelen, J.C. Vorbruggen, B. Sendhoff. Springer, 77-82.

C. Guler, B. Sankur, Y. Kahya, M. Skurichina, S. Raudys (C1996). Classification of respiratory sound patterns by means of cooperative neural networks, EUSIPCO-96, 8. European Signal Processing Conference, September 10-13,1996, Trieste, Italy. Editors G.Ramponi, G.L.Sicuranza, S. Carrato, S.Marsi. isbn 88-86179-83-9. pub.Edizioni Lint Trieste.

S. Raudys (C1995a ). Single layer perceptron and statistical classifiers,- Proc. 3rd Baltic summer school on Mathematical simulation and Informatics, Aug. 1995, KTU Publ. Klaipeda, 41-55  (invited paper).

S. Raudys (C1995 b) Generalization of linear and non-linear adaptive classifiers, Proc. ICANN'95, Paris, October 1995,1, 183-190 (invited paper).

S. Raudys (C1995c ). A negative weight deacay or antiregularization, Proc. ICANN'95, Paris, October 1995, 2, 449-454.

C. Guler, B. Sankur, S. Raudys, Y. Kahya (C1995). Visual classification based on a new mapping method applied to medical diagnosis, 10. Inter. Symp. on Computer & Infor. Sciences, ISCIS-10, ,October 29-31,1995, Izmir, Turkey, 731-738.

S. Raudys (C1994b). Why do ANN classifiers have favorable small sample properties?  Proc. IV-th Int. Conf. Pattern Recognition in Practice, Vlieland, 1-3 June 1994, The Netherlands, 287-298 (invited paper).

S. Raudys (C1994c). Unexpected small sample properties of  perceptrons, Proc. of the First International Conference  Neural Networks and their Applications, Marseilles, France, Dec 15-16, 1994, Publ. IUSPIM, University of Aix-Marseille III, 84-96.

S. Raudys and  V.Diciunas (1994). Generalization error of linear margin classifier,     Proc. 6th microcomputer school. Neural Networks. Theory and    Applications. 18-23 Sept. 1994. Sedmihorky, Czech Republic. Publ. Technical Univ. Brno, 159-164.

.S. Raudys, M. Skurikhina (C1994). Small sample propertieof ridge estimate of the covariance matrix in statistical and neural net classification.  In New Trends in Probability and Statistics, Vol.3, Multivariate Statistics and Matrices in Statistics, Proc. of the 5th Tartu Conference, Tartu-Puhajarve, Estonia, 23-28 May 1994, 237-245.

T. Cibas, F. Fogelman-Soulie, P. Gallinari, S. Raudys (C1994a). Variable selection with optimal cell damage, Proceedings ICANN'94, Sorento, Italy, Marinaro ans P.G. Morso eds. Springer-Verlag, 1, 729-730.

T. Cibas, F. Fogelman Soulie, P. Gallinari, S. Raudys (C1994b). Variable selection with Neural Networks, Proceedings ICANN'94, Sorento, Italy, Marinaro ans P.G. Morso eds. Springer-Verlag, 2 , 1464-1469.

W. Schmidt, S. Raudys, M. Kraaijveld, M. Skurikhina, R. Duin (C1993). Initialisations, back-propagation and generalization of feed-forward ANN classifiers,  Proceedings of  IEEE Conference on  Neural Networks, San Francisko, April 1993, 1,  598-604.

 S. Raudys and M. Skurikhina (C1992). The role of the number of training  samples on weight  initialization of artificial  neural net classifier, Neuroinformatics and Neurocomputers. Proc. RNNS/IEEE Symposium, October, 1992, Rostov-on-Don, Russia. 1992, 343-353.

S. Raudys (C1992). Accuracy of feature selection and extraction in statistical and neural net pattern classification. Proc. of 12th International Conference on Pattern Recognition, September 1992, Hague, The Netherlands, IEEE Computer Society Press, 2, 62-70  (invited paper).

S. Raudys and A. Jain (C1990). Small sample effects in statistical pattern recognition. Recommendations for practitioneers. Proc. of 10-th Int  Joint Conf. on Pattern Recognition, Atlantic City, NJ. USA, June 1990, 417-423.

S. Raudys (C1988). On the accuracy of a bootstrap estimate of the classification error. Proc. of 9-th Int  7th Joint  Conf. on Pattern Recognition, Rome, Italy, Nov. 1988.

S. Raudys (C1987). On the accuracy of model selection in data analysis. Proc of III-rd Int. Conf. on Data Analysis and Informatics , INRIA  Press, Paris, September 1987, 1, 91-100  (invited paper).

S. Raudys, V. Pikelis (C1982c). Accuracy of the best classification algorithm and a subset of variables. Int  6th Joint  Conf. on Pattern Recognition, Munich, Nov. 1982, 19-22.

S. Raudys (C1978). Comparison of the estimates of probability of misclassification. Proc. of 4-th Int  Joint  Conf. on Pattern Recognition, Kyoto, Japan, Nov. 1978, 280-282.

S. Raudys (C1977). On the accuracy of some estimates of the multivariate density function. - Trans. of 7th Prague Conference on Information Theory, Prague: Publ. House of the C zechoslovak Academy of Sciences, A, 429-438.

S. Raudys (C1976). On dimensionality, learning sample size and complexity of classification algorithms. Proc. Third Int. Conf. Pattern Recognition, San Diego, USA,  166-169.

S. Raudys (C1970). On the problems of sample size in pattern recognition - Proc., 2nd All-Union Conf. Statist. Methods in Control Theory; Moscow, Nauka, 2, 64-77.

S.Raudys (C1969). Determination of the realization quantity of recognization  objects for the formation of recognition systems decision rule. - Automatic input of the Written and Printed Characters into Computers, Proc. 2nd All -Union Conf. A. Nasliunas, ed., Vilnius.

References to my students and colleagues papers.

A.D. Deev (1974).  Discriminant function designed on independent blocks of variables.  - Proceedings of Acad. of Sciences of the USSR, Technical Cybernetics, (USSR J.), 12,. 153-156 (in Russian).

K. Juskevicius (1983). Investigation of the sensitivity of a minimum distance piecewise linear classifier to the limitation of the learning sample size. Statistical Methods of Control, 61, 89-129 (in Russian).

L.D. Meshalkin (1976).  Assignement of numerical values to nominal variables.  - Statistical Problems of Control, 14, 49-56 (in Russian).

L.D. Meshalkin V.I. Serdobolskij (1978). Errors in clsssifying multivariate observations.  -Theory of Probabilities and Its Applications , 23(4), 772-781 (in Russian).

V. Pikelis (1973). The errors of a linear classifier with independent measurements when the learning sample size is small. - Statistical Problems of Control, 5 , 69-102 (in Russian).

S. Putinaite (1990). On the expected and apparent probabilities of misclassification of the classifier for discrete variables. - Statistical Problems of Control, 93, 101-111 (in Russian).

M. Skurichina  (1990). Effect of the kernel function form on the quality of nonparametric Parzen window classifier. Statistical Problems of Control, 93, 167-181 (in Russian).

A. Saudargiene (1999). Structurization of the covariance matrix by process type and block-diagonal models in the classifier design. Informatica, 10(2).

A. Raudys (2000). Interactive initialisation of the multilayer perceptron. Pattern Recognition Letters, 21:907-916.

A. Raudys, A. Long (2001). MLP based linear feature extraction for nonlinearly separable data. Pattern Analysis and Applications. 4(4).

Statistical Problems of Contol   were published by Institute of Mathematics and Informatics.