期刊名称：International Journal of Computer Science and Network Solutions
印刷版ISSN：2345-3397
出版年度：2014
卷号：2
期号：8
页码：24-30
出版社：International Journal of Computer Science and Network Solutions
摘要：The relation between the data from Images and other patterns are not explicitly stored in the images and thisleads the image analysis far beyond data mining. Intelligently classification of an image by its content is animportant way to extract meaningful information from a large image collection. One of the most importanttechniques in image mining is automated images classification. This technique involves the basic steps oftraining and testing the system. The system first extracts the basic features of the input images and the dimensionreduction to accelerate system performance using machine learning algorithms to produce a model. This modelis used to classify the test data. For comparing different Classification and performance systems differentevaluation criteria are presented. Among the most important ones accuracy, sensitivity (recall), specificity, Fmeasure,classification error, likelihood ratio and predictive value can be mentioned
