基于机器学习的小容值MLCC容量模型的研究

Research on Small Capacity MLCC Capacity Model Based on Machine Learning

小容值多层瓷介电容器(MLCC)容易出现容量命中率低的问题,本文通过机器学习算法建立小容值MLCC容量模型,采用MSE、R2、拟合精度和预测精度评估各模型的准确度。最后选择采用线性回归模型表达小容值MLCC的印刷层数和电容量之间的关系,并通过实际投产验证,证明该模型可以用于指导小容值MLCC的实际生产,提高小容值MLCC容量命中率。

Small capacitance multilayer ceramic capacitors(MLCC)are prone to the problem of low capacity hit rate.This paper establishes a small capacitance MLCC capacity model using machine learning algorithms,and evaluates the accuracy of each model using MSE,R2,fitting accuracy,and prediction accuracy.Finally,the linear regression model was chosen to express the relationship between the number of printing layers and capacitance of small capacity MLCC,and it was verified through actual production that this model can be used to guide the actual production of small capacity MLCC and improve its capacity hit rate.