Classification of the states of object, which functioning is described by one-dimensional autoregresion
DOI:
https://doi.org/10.34185/1562-9945-5-124-2019-06Abstract
The main purpose of the given work is detection of change of properties of dynamic object that can be in two classes of states, and in each of them the behaviour of object is described by the own autoregression model with unknown parameters.
The problem of detection of change of properties of dynamic objects frequently arises in such areas, as technical and medical diagnostics, the control of technological processes, monitoring, processing of signals.
It is supposed, that there are two groups of observations of functioning of object, which are described by two different autoregression models. Basing on the decision of tasks of identification of two autoregression models, it is necessary to construct deciding rule, which would allow relating new observations to one of two groups of observations. It is supposed, that structures of two models are known, and covariance matrixes of random component of supervision of output variable in two conditions of object can be various.
Basing on the original results on identification autoregression models, the author successfully solves the task of statistical classification. The rule of statistical classification is constructed and its properties are investigated.
References
Obnaruzhenie izmeneniya svoystv signalov i dinamicheskih sistem: per. s angl. / M. Bassvil, A. Vilski, A. Banvenist i dr.; pod. red. M. Bassvil, A. Banvenista. – M. : Mir, 1989. – 278 s.
Borodkin L. I. Algoritm obnaruzheniya momentov izmeneniya parametrov uravneniya sluchaynogo protsessa / L. I. Borodkin, V. V. Mottl // Avtomatika i telemehanika. – 1976. – № 6. – S. 23–31.
Kartashov V. Ya. Obnaruzhenie strukturno-parametricheskih izmeneniy v stohasticheskih sistemah v realnom masshtabe vremeni algoritmami nepreryivnyih drobey i strukturnogo analiza / V. Ya. Kartashov, M. A. Novoseltseva // Upravlenie bolshimi sistemami. – 2011. – Vyip. 34. –
S. 62–91.
Mirgorod V. F. Matematicheskie modeli protsessov upravlyaemogo izmeneniya sostoyaniya silovyih i energeticheskih ustanovok : diss. … d-ra tehn. nauk / Mirgorod Vladimir Fуodorovich. – Dnepropetrovsk, 2012. – 344 s.
Sarychev A. P. Lineynaya avtoregressiya na osnove metoda gruppovogo uchyota argumentov v usloviyah kvazipovtornyih nablyudeniy / A. P. Sarychev // Iskusstvennyiy intellekt. – 2015. – № 3–4 (69–70). – S. 105–123.
Sarychev A. P. Identifikatsiya parametrov sistem avtoregressionnyih uravneniy pri izvestnyih kovariatsionnyih matritsah / A. P. Sarychev // Mezhdunarodnyiy nauchno-tehnicheskiy zhurnal “Problemyi upravleniya i informatiki”. – 2012. – № 3. – S. 14–30.
Sarychev A. P. Issledovanie metodom statisticheskih ispyitaniy iteratsionnoy protseduryi dlya identifikatsii parametrov sistemyi avtoregressionnyih uravneniy / A. P. Sarychev // SistemnI tehnologiei. –2014. – Vipusk 3 (92). – S. 77–89.
Seber Dzh. Lineynyiy regressionnyiy analiz : per. s angl. / Dzh. Seber. – M. : Mir, 1980. – 456 s.
Anderson T. Vvedenie v mnogomernyiy statisticheskiy analiz : per. s angl. / T. Anderson. – M. : Fizmatgiz. – 1963. – 500 s.
Spravochnik po teorii veroyatnostey i matematicheskoy statistike / V. S. Korolyuk, N. I. Portenko, A. V. Skorohod, A. F. Turbin. – M. : Nauka, 1985. – 640 s.
Sarychev A. P. Modelirovanie v klasse sistem avtoregressionnyih uravneniy v usloviyah strukturnoy neopredelennosti / A. P. Sarychev // Mezhdunarodnyiy nauchno-tehnicheskiy zhurnal “Problemyi upravleniya i informatiki”. – 2015. – № 4. – S. 79–103.
Aleksandr Sarychev Modelirovanie slozhnyih sistem v usloviyah strukturnoy neopredelyonnosti: regressionnyie i avtoregressionnyie modeli / A. P. Sarychev LAP LAMBERT Academic Publishing RU, Saarbrücken, Deutschland. – 2016. – 274 с.
