Original scientific paper

Regression analysis and neural networks as methods for production time estimation

Predrag Ćosić ; University of Zagreb, Faculty of Mechanical Engineering and Naval Architecture, Ivana Lucića 5, 10000 Zagreb, Croatia
Dragutin Lisjak orcid.org/0000-0002-9976-577X ; University of Zagreb, Faculty of Mechanical Engineering and Naval Architecture, Ivana Lucića 5, 10000 Zagreb, Croatia
Dražen Antolić ; ADPP, Ilirski trg 5, 10000 Zagreb, Croatia

Abstract

An experienced process planner usually makes decisions based on comprehensive data without breaking it down into individual parameters. So, as the first phase it was necessary to establish a technological knowledge base, define features of the 2D drawing (independent variables), possible dependent variables, size and criteria for sample homogenization (principles of group technology) for carrying out analysis of variance and regression analysis. The second phase in the research was to investigate the possibility for easy automatic, direct finding and applying 3D features of an axial symmetric product to the regression model. The third phase in the research was to investigate the possibility for the application of neural networks in production time estimation and to compare the 224 results between the regression models and neural network models. The most important characteristic of our approach presented in this paper is estimation of production times using group technology, regression analysis and neural networks.

Keywords

group technology; knowledge base; neural networks; production time; stepwise multiple linear regression; Total Cost Estimation

Hrčak ID:

75393

URI

https://hrcak.srce.hr/75393

Publication date:

27.12.2011.

Article data in other languages: croatian

Visits: 3.834 *