Open AccessUżyteczność modeli parametrycznych i sztucznych sieci neuronowych w prognozowaniu kosztów produkcji
Author(s) -
Zbigniew Leszczyński,
Tomasz Jasiński
Publication year - 2017
Publication title -
zeszyty teoretyczne rachunkowości
Language(s) - English
Resource type - Journals
eISSN - 2391-677X
pISSN - 1641-4381
DOI - 10.5604/01.3001.0009.8025
Subject(s) - production (economics) , parametric statistics , artificial neural network , nonparametric statistics , estimation , cost estimate , computer science , dependency (uml) , process (computing) , operations research , econometrics , engineering , economics , artificial intelligence , mathematics , statistics , systems engineering , microeconomics , operating system
Użyteczność modeli parametrycznych i sztucznych sieci neuronowych w prognozowaniu kosztów produkcjiThe aim of the paper is to analyze parametric models and artificial neural networks in terms of their suitability as estimation tools of the production costs. Estimated production costs are a fundamental determinant of the decision-making process by costs engineers relating to design and management costs of new products, infrastructure projects and production lines. The first part of the paper presents a con- ceptual framework for the construction of a model of production costs parametric estimation, multidimensional with linear and nonlinear dependency. It then discusses the nature and use of artificial neural networks as nonparametric estimates of production costs. In both parts of the article, an empirical study is conducted with the use of adequate statistical methods and artificial neurons. This study presents proce- dures for construction of models of parametric and nonparametric estimation of production costs and discusses their advantages and disadvantages. It also presents the application and usefulness of both models for estimating production costs in production environment