| 摘要 |
<p><P>PROBLEM TO BE SOLVED: To estimate the degree of state of a feature (for example, disease, abnormality, etc.) of a test sample even if the feature is unknown. <P>SOLUTION: A featureless standard sample and a sample having a first feature are selected as control samples of a first class, and the standard sample and a sample having a second feature are selected as control samples of a second class. SIMCA analysis is performed by using the control samples of the first and second classes and test samples, and a distance of each sample from each of the first and second classes is calculated, and a Coomans plot in which all the samples are plotted is formed and displayed. The state degree of each test sample is estimated from the degree of state of the control sample having a coordinate close to the coordinate of each test sample on the Coomans plot. <P>COPYRIGHT: (C)2009,JPO&INPIT</p> |
