| 摘要 |
<p>PURPOSE:To provide the inferred result more reducing error by complementing the faults of forward and backward diagnostic processing by performing prescribed weighting to the respective degrees of certainty of inference due to two inferring means and finally inferring the degree of certainty for each of various causes based on these respective weighted degrees of certainty. CONSTITUTION:In a step S101, the present degree of abnormality of a phenomenon is inputted. In S102, the degree of certainty for each cause is inferred based on generation frequency data and relation degree data. In S103, as a second inferring means (backward diagnostic processing), predictive generation frequency data are compared with the generation frequency data for each cause, and the degree of certainty for each cause is inferred so that the more the error of the cause is reduced, the more the degree of certainty can be improved. In S104, prescribed weighting is performed to the respective degrees of certainty discussed at the S102 and S103, and the degree of certainty for each cause is finally inferred based on these respective weighted degrees of certainty.</p> |
