01.00.00 Physical-mathematical sciences
-
NONPARAMETRIC AND APPLIED STATISTICS IN OUR COUNTRY
01.00.00 Physical-mathematical sciences
DescriptionWe continue the series of articles about the history of statistics. We discuss the development of nonparametric and applied statistics in our country in 1930 - 1980 years. We have presented the studies of the great statisticians of the twentieth century, such as N.V. Smirnov, L.N. Bolshev, V.V. Nalimov. American statistics show Russian debate about applied statistics. We have briefly listed the process of creation of the All-Union Statistical Association (1990) and its further developments
-
01.00.00 Physical-mathematical sciences
DescriptionIn the article we have considered A. N. Kolmogorov and N. V. Smirnov papers dedicated to one-sided and two-sided goodness-of-fit and homogeneity tests. It has been shown that the term "Kolmogorov - Smirnov test" used incorrectly. We have also given the recommendations on use of the terms "Kolmogorov test", "Smirnov test", "test of Kolmogorov-Smirnov type" and discussed omega-square test (Cramer-von Mises–Smirnov test). Typical errors in the application of these criterions have been considered, in particular to test for normality of the distribution of measurement results
-
NON-PARAMETRIC CYCLES ESTIMATORS
01.00.00 Physical-mathematical sciences
DescriptionIn many applications, we study the time series (or a random process), which is the sum of the periodic deterministic function of time and random errors that distort the periodic signal. It is required to estimate the length of the period and the periodic component. It does not assume that the periodic function is included in any parameter family of functions, such as finite sums of sines and cosines. It is obvious that the assumption of occurrence of a periodic function in parametric family does not meet the characteristics of the real world, ie, is conditional, internal mathematical (look for the keys under the lamp because there is a light, not in the bush, where lost, because there are dark). For similar reasons, it is impossible to assume that the distribution function of the random errors is included in any parameter family of distributions. In accordance with the new paradigm of mathematical statistics in this article we studied the problem of nonparametric estimation (minimum) length of the period and the periodic component of the signal. On the basis of natural variation and scope of indicators is suggested a new class of nonparametric estimators of the length of the period and the periodic component in the time series. Based on the general results of statistics of objects of non-numeric nature we proved the consistency of these estimates. From the practical point of view it is necessary to minimize the numerical (one parameter - ability length of period of time) one or more of the 66 functionals, described in the article
-
NONPARAMETRIC KERNEL ESTIMATORS OF PROBABILITY DENSITY IN THE DISCRETE SPACES
01.00.00 Physical-mathematical sciences
DescriptionSome estimators of the probability density function in spaces of arbitrary nature are used for various tasks in statistics of non-numerical data. Systematic exposition of the theory of such estimators has been started in our articles [3, 4]. This article is a direct continuation of these works [3, 4]. We will regularly use references to conditions and theorems of the articles [3, 4], in which introduced several types of nonparametric estimators of the probability density. We have studied linear estimators. In this article, we consider particular cases - kernel density estimates in discrete spaces. When estimating the density of the one-dimensional random variable, kernel estimators become the Parzen-Rosenblatt estimators. Under different conditions, we prove the consistency and asymptotic normality of kernel density estimators. We have introduced the concept of "preferred rate differences" and are studied nuclear density estimators based on it. We have introduced and studied natural affinity measures which are used in the analysis of the asymptotic behavior of kernel density estimators. Kernel density estimates are considered for sequences of spaces with measures. We give the conditions under which the difference between the densities of probability distributions and of the mathematical expectations of their nuclear estimates uniformly tends to 0. Is established the uniform convergence of the variances. We find the conditions on the kernel functions, in which take place these theorems about uniform convergence. As examples, there are considered the spaces of fuzzy subsets of finite sets and the spaces of all subsets of finite sets. We give the condition to support the use of kernel density estimation in finite spaces. We discuss the counterexample of space of rankings in which the application of kernel density estimators can not be correct
-
NONPARAMETRIC ESTIMATION OF CHARACTERISTICS OF PROBABILITY DISTRIBUTIONS
01.00.00 Physical-mathematical sciences
DescriptionThe article is devoted to the nonparametric point and interval estimation of the characteristics of the probabilistic distribution (the expectation, median, variance, standard deviation, variation coefficient) of the sample results. Sample values are regarded as the implementation of independent and identically distributed random variables with an arbitrary distribution function having the desired number of moments. Nonparametric analysis procedures are compared with the parametric procedures, based on the assumption that the sample values have a normal distribution. Point estimators are constructed in the obvious way - using sample analogs of the theoretical characteristics. Interval estimators are based on asymptotic normality of sample moments and functions from them. Nonparametric asymptotic confidence intervals are obtained through the use of special output technology of the asymptotic relations of Applied Statistics. In the first step this technology uses the multidimensional central limit theorem, applied to the sums of vectors whose coordinates are the degrees of initial random variables. The second step is the conversion limit multivariate normal vector to obtain the interest of researcher vector. At the same considerations we have used linearization and discarded infinitesimal quantities. The third step - a rigorous justification of the results on the asymptotic standard for mathematical and statistical reasoning level. It is usually necessary to use the necessary and sufficient conditions for the inheritance of convergence. This article contains 10 numerical examples. Initial data - information about an operating time of 50 cutting tools to the limit state. Using the methods developed on the assumption of normal distribution, it can lead to noticeably distorted conclusions in a situation where the normality hypothesis failed. Practical recommendations are: for the analysis of real data we should use nonparametric confidence limits
-
01.00.00 Physical-mathematical sciences
DescriptionAn analysis of the experimental data obtained by the authors, as well as reference books, allowed to hypothesize about the essential role of gravitational convection in electromembrane systems with ampholytes even in underlimiting current regimes. The article is devoted to the development of the mathematical model of ion transport in a flow elecrtomembrane system during electrodialysis of ampholyte-containing solutions with taking into account a possible appearance of gravitational convection, in particular, due to nonisothermal protonation–deprotonation reactions of ampholytes. The article presents the boundary value problem that is the new mathematical model for diffusion, convection and electromigration of four components of the solution (ions of sodium, dihydrogen phosphate and hydrogen, as well as molecules of orthophosphoric acid) in a half of an electrodialysis desalination channel, adjacent to an anion-exchange membrane. The membrane is considered as ideally selective and homogeneous. The system of partial differential equations, that is the base of the model, also includes equations of Navier-Stokes, material balance, convective heat conduction and the electroneutrality condition. The system of equations is supplemented by a number of natural and original boundary conditions. A distinctive feature of this study is the absence of assumptions about the equilibrium of chemical reactions in a diffusion layer. The results of the study can be used for the development of environmentally rational and resource saving membrane technologies for a processing of products of agro-industrial complex
-
01.00.00 Physical-mathematical sciences
DescriptionClassic quantitative measure of the reliability of the models: F-measure by van Rijsbergen is based on counting the total number of correctly and incorrectly classified and not classified objects in the training sample. In multiclass classification systems, the facility can simultaneously apply to multiple classes. Accordingly, when the synthesis of the model description is used for formation of generalized images of many of the classes it belongs to. When using the model for classification, it is determined by the degree of similarity or divergence of the object with all classes, and a true-positive decision may be the membership of the object to several classes. The result of this classification may be that the object is not just rightly or wrongly relates or does not relate to different classes, both in the classical F-measure, but rightly or wrongly relates or does not relate to them in varying degrees. However, the classic F-measure does not count the fact that the object may in fact simultaneously belongs to multiple classes (multicrossover) and the fact that the classification result can be obtained with a different degree of similarity-differences of object classes (blurring). In the numerical example, the author states that with true-positive and true-negative decisions, the module similarities-differences of the object classes are much higher than for false-positive and false-negative decisions. It would therefore be rational to the extent that the reliability of the model to take into account not just the fact of true or false positive or negative decisions, but also to take into account the degree of confidence of the classifier in these decisions. In the intellectual system called "Eidos", which is a software toolkit for the automated system-cognitive analysis (ASC-analysis), we use initially proposed by its developers measure of the reliability of the models, which is essentially a fuzzy multiclass generalization of the classical F-measure (it is proposed to call it the L-measure). In this article, L-measure is mathematically described and its application is demonstrated on a simple numerical example
-
01.00.00 Physical-mathematical sciences
DescriptionThe relationship of Mathematical Statistics (wider - Mathematical methods of research) and history is multifaceted. In our opinion, the history of mathematical statistics is an integral part of this mathematical discipline. We have given a review of our works on the history of statistical methods. The role of mathematical statistics for the history is very important. In this article, we restrict ourselves to the questions of chronology. For centuries, the chronology is considered as a part of applied mathematics. The main problem is that the whole "common" concept of the Russian and the World history as a whole presented in textbooks was faked by the opponents of Russia after the collapse of the global Empire (Russian kingdom) in the early 17th century - 400 years ago. The stories about historical events are the information weapon. It was used by the new rulers to suppress the resistance of the vanquished. A new mathematical and statistical chronology of general and Russian history, which was built by a scientific team led by Academician Fomenko, has been helpful for the discussion about the current economic and political problems of relations between Russia and the West in the XXI century. In our opinion, the new chronology of the World and Russian history should be one of the foundations of state-patriotic ideology and deriving practical solutions. The purpose of this article is to give the initial idea of the new chronology from this point of view
-
A NEW APPROACH TO THE STUDY OF THE STABILITY OF CONCLUSIONS IN THE MATHEMATICAL MODELS
01.00.00 Physical-mathematical sciences
DescriptionWe have proposed the general scheme for studying the stability of the conclusions obtained by mathematical methods and models regarding tolerance deviations of the original data and background models. The concrete problems of stability are discussed: towards a change of data, its size and distributions, to allowable transformations measurement scales, to the temporal characteristics (time of start of the project, the planning horizon). Reducing the uncertainty can be conducted by changing the type of data, i.e. with the aid of the transition to non-numerical data. The models of concrete management processes of industrial organizations are considered on the examples of stability characteristics of investment projects to change the discount factors and in models of inventory management to change in the coefficients of the model and batch size production
-
THE INTERACTION OF LIGHT AND PARTICLES WITH GRAVITATIONAL WAVES
01.00.00 Physical-mathematical sciences
DescriptionIn this article we consider a model of the structure of matter, in which elementary particles, atoms and molecules are composed of gravitational waves. A model of interaction of light and particle beams with macroscopic gravitational waves has been proposed. The protocols of experiments to test the theory are considered