Research on Model of Circuit Fault Classification Based on Rough Sets and SVM

摘要

Aiming at the characteristic of lacking swatches and paroxysmal faults, A fault classification model based on rough sets and SVM is put forward. The pretreatment of diagnosis data is constructed by attribute reduction in rough sets. Redundancy attribute is deleted from the diagnosis decision-making table without losing useful information, and the reduced diagnosis decision-making table is used as original training sets of classification sub-system. The dimension of fault symptom and the capability of classification is balanced. Finally an example shows the model is effective and reasonable.