| 摘要 |
A method and a device for measuring the performance of a diagnoser of a system under diagnosis, which comprises a knowledge base comprising a set of failures of the system, of effects observed and of relations between failures and effects observed, and an algorithmic core. The method comprises a step of calculating a specific detection rate and a specific location rate for the knowledge base, respectively referred to as expected detection rate and expected location rate. It also comprises a step of calculating a specific detection rate and a specific location rate for the set consisting of knowledge base and algorithmic core, respectively referred to as detection rate obtained and location rate obtained; and a step of calculating the performance of the diagnoser as a function of the expected and obtained detection rates and of the expected and obtained location rates. |
