| 摘要 |
PROBLEM TO BE SOLVED: To diagnose whether an examinee is suffering from any diseases, kinds of the diseases from which he/she is suffering, the degree or possibility of the disease, attributes of the examinee, and the degree or possibility of the attributes from waveforms obtained from the object to be examined with high accuracy. SOLUTION: Equipment is provided with the following parts: A computing part 13 for components of high frequency computes the components of high frequency up to a resolution by applying wavelet transform on signals from the examined object measured as potential waveforms. A part 14 for producing waveform vectors produces vectors of a certain dimension by summing squares of components of high frequency. A part 15 for determining groups divides vectors of a certain dimension into plural groups using a threshold value and directs the object to be diagnosed to a suitable group. A part 16 for making a vector for the main component makes a vector for the main component by extending the vector of a certain dimension to the group where it belongs with a characteristic vector obtained by the analysis of the main component. A diagnosing part 17 has a three-layered neural network which inputs the vector of the main components, and it retains the number assigned to the output unit which is found to have the largest output value. |
