METHODOLOGY OF THE INFORMATIVE FEATURES SELECTION FOR THE OBJECT CLASSIFICATION BASED ON THE PRINCIPAL COMPONENT ANALYSIS
Authors:
1.
aint-Petersburg state university of information technologies, mechanics and optics
2.
Siberian fire and rescue academy - branch of Saint-Petersburg university of State fire service of EMERCOM of Russia
3.
Siberian fire and rescue academy - branch of Saint-Petersburg university of State fire service of EMERCOM of Russia
Type:
Article
Pages:
from 40
to 44
Status:
Published
Received:
12.09.2014
Accepted:
20.09.2014
Published:
25.09.2014
Language:
Russian
Keywords:
bayesian approach, object classification, cluster analysis, fire safety, fire risk assessment
Abstract (English):
The possibility of using the Bayesian approach for assessing the fire hazard of buildings and structures is considered. The method of feature selection is offered not only for solving the problem of the object classification, but also for the cluster analysis. That will allow to provide the minimum average risk of erroneous decisions.
The possibility of using the Bayesian approach for assessing the fire hazard of buildings and structures is considered. The method of feature selection is offered not only for solving the problem of the object classification, but also for the cluster analysis. That will allow to provide the minimum average risk of erroneous decisions.
Keywords:
bayesian approach, object classification, cluster analysis, fire safety, fire risk assessment
bayesian approach, object classification, cluster analysis, fire safety, fire risk assessment
1. Metodika opredeleniya raschetnyh velichin pozharnogo riska v zdaniyah, sooruzheniyah i stroeniyah razlichnyh klassov funkcional'noy pozharnoy opasnosti: Prilozhenie k Prikazu MChS Rossii ot 30 iyunya 2009 g. № 382 // Elektronnyy fond ravovoy i normativno-tehnicheskoy dokumentacii. URL: http//www.docs.cntd.ru (data obrascheniya: 22.05.2014).
2. Fukunaga K. Vvedenie v statisticheskuyu teoriyu raspoznavaniya obrazov. M.: Nauka, 1979. 346 c.
3. Kahaner D., Mouler K., Nesh S. Chislennye metody i programmnoe obespechenie: per. s angl.; pod red. H.D. Ikramova. M.: Mir, 1998. 575 s.
