Julius Kruopis - taikomosios statistikos pradininkas Lietuvoje

LTStraipsnyje apžvelgta Lietuvos taikomosios statistikos pradininko Juliaus Kruopio mokslinė ir pedagoginė veikla, parengti vadovėliai ir monografijos. Plačiausiai aprašytas jo indėlis į statistikos mokslą, bendradarbiavimas su Lietuvos įmonėmis taikant statistinius metodus įvairiose žmogaus veiklos srityse. Paminėti ir J. Kruopio darbai iš tikimybių teorijos srities. Aptarti jo pedagoginės veiklos ypatumai, vadovavimas studentų baigiamiesiems darbams ir disertacijoms. Raktiniai žodžiai: Julius Kruopis; matematinė statistika; statistikos taikymai; kokybės kontrolė; Puasono aproksimacija. [Iš leidinio]

ENJulius Kruopis was born on 21 February 1941 in Utena district. In 1963 he graduated from Vilnius University, Faculty of Physics and Mathematics. In 1964–1966, he worked as a trainee lecturer at the Department of Probability Theory and Number Theory of the Faculty of Mathematics of Vilnius University. In 1966–1969, he studied at the post-graduate course of the Steklov Institute of Mathematics of the USSR (his supervisor was the famous Russian mathematician Professor Login Bolshev) and defended the thesis of a candidate of mathematical sciences (the current equivalent of PhD thesis). After returning to Lithuania after postgraduate studies, J. Kruopis worked as a senior lecturer at the Department of Probability Theory and Number Theory of the Faculty of Mathematics of Vilnius University. In 1971, he was transferred to the newly established Department of Applied Mathematics (from 1972, he worked as an associate professor). In 1991, this department was renamed as the Department of Mathematical Statistics. Since 2013, J. Kruopis has been an affiliated professor of the Department of Mathematical Statistics. In 1980–1986, J. Kruopis headed the Department of Applied Mathematics. He established and fostered new traditions and aimed for the majority of students’ theses to have a clear applied nature. He led such works himself and encouraged other teachers to do the same. From the very first days of scientific work, the talent of J. Kruopis became evident through mathematical and statistical modeling of real processes. PhD thesis “Estimation of Markov processes parameters” by J. Kruopis, defended in 1969 in Moscow (supervised by prof. L. Bolshev), has a clear applied character. Statistical modeling of tick-borne encephalitis virus infection and immunity formation processes was performed.Methods of queueing theory and mathematical statistics were used for modeling and statistical estimation of model parameters. A detailed analysis of the obtained results was carried out using real data. In addition, the thesis presents the assessment of the transition probabilities of Markov processes with migration. The findings are used for demographic and reliability data analysis. After returning to Lithuania, J. Kruopis developed a very wide range of activities in the application of statistical methods in many companies and factories. He consulted engineers of Kaunas radio factory, Panevėžys factory of kinescopes “Ekranas” and joint stock company “AB Lietkabelis”, Vilnius drill, “Venta”, “Vilma”, radio components factories, Šiauliai TV factory, Dniepropetrovsk tire factory “Dnieproshina”, Moscow Aviation Institute, etc. Julius Kruopis grasped the essence of technological processes very quickly. A particularly long and fruitful cooperation was established with the Kaunas radio plant. Under the leadership of J. Kruopis, a number of factory managers and engineers defended PhD theses. His works on the application of statistical methods are especially valuable. These applications are mainly related to optimization of control systems, regulation and management of technological processes (acceptance control of materials and details, intermediate control of manufactured details and nodes, outgoing production control, metrological control of measuring devices, product reliability control, etc.). Although his work is predominated by works on statistical applications, there are also theoretical works on mathematical statistics. For example, the problem of making an optimal decision when choosing a distribution from several possible alternative distributions was examined.There are several works by J. Kruopis related to the application of statistical methods in reliability theory and survival analysis. Although J. Kruopis always considered himself primarily a statistician, he also achieved important results in probability theory. In 1986, J. Kruopis introduced shifted Poisson approximation, which, unlike the usual Poisson approximation has two parameters and, therefore, is comparable to normal approximation. Moreover, for integer-valued random variables, the accuracy of shifted Poisson approximation can be measured in total variation metric. In contrast, the accuracy of normal approximation to discrete random variables is measured in weaker uniform metric only. A large part of J. Kruopis research in Probability Theory is related to so-called Signed Compound Poisson approximations – a topic which was quite new in the early eighties. J. Kruopis introduced so-called left-hand-side and right-hand-side factorial cumulants and practically solved the problem of the choice of parameters for such approximations. J. Kruopis is the co-author of several monographs: “Quality of mechatronic products. Selective control”, “Nonparametric tests for complete data”, “Nonparametric tests for censored data”. While working at the university, J. Kruopis supervised a large number of undergraduate, bachelor, and master theses. Under his supervision Vanda Bikelienė (Asymptotic distribution of sums of samples from finite populations, 1981) and Natalja Kosareva (Asymptotic analysis of statistics distributions in the scheme of sets of particles arrangement with equal probabilities, 1986) defended their doctoral theses in mathematics. [...] Keywords: Julius Kruopis; mathematical statistics; statistical applications; quality control; Poisson approximation. [From the publication]