During planning and designing of new products the design managers are interested in estimating the cost as early as possible. However, at the early design stage only a few attributes of the future product are known, and their impact on cost is not clear to the cost estimation expert. Neural networks can be used to detect the hidden relationships between cost drivers and the cost of a new product, and estimate the cost after being presented a small set of conceptual attributes describing the product. Based on a laboratory benchmark example with artificial data we present our experiments on classification perceptrons with one hidden layer. The attention is focused on the problem of small number of training samples available in the domain. The results of three possible remedies are presented namely: prewiring background knowledge, preprocessing input data, and transforming input data.
[1]
Hans Holm,et al.
Computer Applications in Production and Engineering.
,
1989
.
[2]
S. Schaal,et al.
In CAD integrierte Kostenkalkulation
,
1992
.
[3]
Jürgen Bode,et al.
Neural networks in new product development
,
1995
.
[4]
Nenad Ivezic,et al.
A neural network-based machine learning approach for supporting synthesis
,
1994,
Artif. Intell. Eng. Des. Anal. Manuf..
[5]
Ivan Bratko,et al.
Machine learning in artificial intelligence
,
1993,
Artif. Intell. Eng..
[6]
Jörg Becker,et al.
Supporting the Design Process with Neural Networks
,
1993
.
[7]
Jürgen Bode,et al.
Cost estimation in new product development: Knowledge pluralism, knowledge granularity, and agent based representation
,
1995
.