Analog Circuit Fault Diagnosis based on Optimization Matrix Random Forest Algorithm

Abstract

Aiming at the existing artificial neural network, support vector machine and other artificial intelligence algorithms for analog circuit fault diagnosis, and algorithms need to be long and varied. This paper proposes a random forest optimization matrix model algorithm for analog circuit fault diagnosis. The method based on random forest algorithm, on the basis of through three incentive to establish a special optimization matrix model. When the circuit failure occurs, as the excitation input and output response matrix elements in the change. According to the optimization of matrix model, and the characteristics of random forest algorithm, using multidimensional vector can have different effective characteristics. The optimization of matrix model is combined with bagging and decision trees, can accurate single fault and multiple faults of analog circuit fault diagnosis research. Compared with other types of artificial intelligence algorithms, the optimized matrix random forest algorithm can meet the requirements of both feature extraction and effective classification. And the fault diagnosis rate reaches 99.5%.