THE METHOD OF MIXED ESTIMATION OF REGRESSION MODEL PARAMETERS WITH AN ARBITRARY
COMPOSITION OF OBSERVATION GROUPS 

Sergey I. Noskov, Doctor of technical sciences, professor, professor of the sub-department of information systems and information security, Irkutsk State Transport University (15 Chernyshevskogo street, Irkutsk, Russia) sergey.noskov.57@mail.ru
Egor S. Popov, Master degree student of the sub-department of information systems and information security, Irkutsk State Transport University (15 Chernyshevskogo street, Irkutsk, Russia) eglir5732@mail.ru

Background. The construction of regression models of various complex objects in accordance with the objectives of the study is sometimes accompanied by the need to split the original data sample into subsamples (groups of observations) or the use of the so-called panel data. The purpose of the study is to modify the method for calculating estimates of the parameters of a linear regression equation developed earlier by one of the authors, using the mixed estimation method. Materials and methods. To achieve this goal, the mathematical apparatus for solving linear and linear Boolean programming problems was used. Results. This goal is formalized by setting the task of automating the formation of the composition of the index set of observation numbers to implement the antirobust estimation method within the framework of the mixed estimation method by solving a linear Boolean programming problem. Conclusions. The approach described in the work allows us to combine the advantages of the least modulus and anti-robust estimation methods when modeling. A regression model of freight turnover of road transport in the Russian Federation has been constructed that is adequate to the analyzed object.

regression model, parameters estimation methods, mixed estimation method, linear and linear Boolean programming problems, road transport freight turnover

Noskov S.I., Popov E.S. The method of mixed estimation of regression model parameters with an arbitrary composition of observation groups. Modeli, sistemy, seti v ekonomike, tekhnike, prirode i obshchestve = Models, systems, networks in economics, technology, nature and society. 2024;(3):98–104. (In Russ.). doi: 10.21685/2227-8486-2024-3-8