PCB Characteristic Impedance Prediction based on an Error Compensated Random Forest Regression Model

The characteristic impedance is a key factor affecting the quality of signal transmission in PCB (Printed Circuit Board). In order to solve the problem of characteristic impedance prediction during the PCB manufacturing process, a random forest regression model with support vector regression as error compensation is developed. The performance of the presented method is evaluated on the data from real manufacturing process. The obtained results were rather satisfied compared with the traditional prediction model. The R-square is 0.99 and RMSE is 0.12 which would be useful for the real application.