The application and research of fault detection based on PC-KNN in semiconductor batch process

In this paper, PC-KNN is studied on the condition that the data dimension is reduced by PCA. FD-KNN (Fault Detection based on K - Nearest - Neighbor) has been applied in semiconductor manufacturing Fault Detection, it can handle nonlinear and multi modal testing problems that influence the performance of PCA. The computational complexity and higher requirement of time and storage space have become the major factors which influence performance of FD-KNN. First, PCA is used to reduce the dimension of original data, then FD-KNN method is applied in principal space, it can effectively reduce the complexity of the calculation and the requirements of system resources process. Through the application in semiconductor batch production process, the results show the performance of PC-KNN dealing with nonlinear and multimodal , it demonstrate the effectiveness of the method proposed in this paper.