ART-based Fuzzy Classifiers: ART Fuzzy Networks for Automatic Classification

Vorschau

Inhaltsverzeichnis, Datei (130 KB)
Leseprobe, Datei (91 KB)

Beschreibung

Quality testing plays an important role in modern industry, producing a huge amount of elaborate technical products. Due to the increasing complexity of design and the high automation of technological processes it is becoming more crucial for industrial companies than it was in past decades. Constantly rising demands on product quality, cost pressure from customers, and the increased prices of raw materials are forcing manufacturers to explore novel ways of quality testing. The objective is to monitor in-process data as a means of indicating product quality and to react to unexpected technological disturbances as soon as possible. If a problem is not recognized quickly, many defective devices will be produced in the meantime. More and more complicated technology has resulted in more severe requirements for quality testing, even as it has caused an escalation of the information stream to be processed. Statistical methods used in the branch earlier cannot provide adequate speed and quality of data analysis in response to a continuously strengthened load. Another limitation of these tools is a lack of automation, adaptive behavior and flexibility, which has to be compensated by involving additional human resources. However, a hot trend in industry is the development of diagnostic systems with minimum functions for an operator. Therefore the need for automatic and intelligent methods of data analysis has arisen and seems likely to become ubiquitous in the near future.