Colored information from a black box?: Validation and evaluation of neural networks
โ Scribed by G. Kateman; J.R.M. Smits
- Book ID
- 102984273
- Publisher
- Elsevier Science
- Year
- 1993
- Tongue
- English
- Weight
- 946 KB
- Volume
- 277
- Category
- Article
- ISSN
- 0003-2670
No coin nor oath required. For personal study only.
โฆ Synopsis
Artificial neural networks are used in many fields and have proven their worth. Until recently the applications in analytical chemistry are scarce, but now many groups are working in this field. One of the reasons for the hesitation in accepting artificial neural networks is the weak scientific understanding of the system. In fact it works as a black box. However, the applications in analytical chemistry so far are so encouraging that more research will be conducted. One very important aspect will be the critical validation of artificial neural networks as long as no sound theoretical background is available. Some criteria will be discussed and some practical examples will be given.
๐ SIMILAR VOLUMES
## Abstract Proposed in this article are two kinds of emotional models based on the neural network and the adaptive fuzzy system that can transform the physical features of a color pattern into its emotional features. The purpose of this system of models is to evaluate the neural network and adapti