| 摘要 |
PURPOSE:To easily perform high-accuracy inference in a short time when changing the degrees of certainty for various causes by calculating the predictive generation frequencies of respective phenomena for each degree of certainty for each of these causes based on predictive generation frequency data. CONSTITUTION:In a step S101, a present abnormal degree Ii of a phenomenon (i) is inputted. In S102, based on the predictive generation frequency data, a predictive abnormal degree Pij(k) as the predictive generation frequency of each phenomenon (i) assuming the degree of certainty for a cause (j) is (k) is calculated for each degree (k) of certainty for various causes (j) based on the predictive generation frequency data. In S103, error Eij (k) of the abnormal degree Ii is calculated for each degree of certainty for various causes by comparing the predictive abnormal degree Pij(k) for each degree of certainty for each of these respective causes (j) with the generation degree data. In S104, a level ERRORj(k) of error Eij(k) is calculated for each cause (j) and in S105, a final degree CFj of certainty is calculated from the distribution of this level ERRORj(k). In S106, the diagnosed result is displayed. |
