| 摘要 |
PROBLEM TO BE SOLVED: To provide a system for simply discriminating the morbid state of hepatitis with high precision using the examination result of blood by solving the problem that a doctor estimates a morbid state from the examination result of blood heretofore in order to discriminate the morbid state of a C-type hepatitis patient but this discrimination method is not practical because discrimination precision is about 70% while, in the case of diagnosis using the biopsy of a liver, there is problem in labor taken in examination or dangerousness. SOLUTION: A distinctive feature (examination item) effective for discriminating the morbid state of hepatitis is extracted from the blood examination data of the C-type hepatitis patient by a feature selecting algorithm SFFS (Sequential Forward Floating Search) and the morbid state is discriminated by SVM (Support Vector Machine) said to be one of most superior learning methods in pattern recognition using it. The morbid state is discriminated in a precision of 94% by the above-mentioned system. COPYRIGHT: (C)2006,JPO&NCIPI
|
