虚拟仪器非线性校正的支持向量机方法

Support vector machine approach to nonlinear calibration for virtual instruments

随着虚拟仪器的广泛应用,其误差控制问题越来越突出。传统的虚拟仪器非线性校正主要采用人工神经网络的方法,由于该方法本身固有的缺陷,其应用受到一定限制。支持向量机是近年来发展起来的一种新的机器学习算法,在许多领域中得到应用。本文分析了虚拟仪器的非线性误差的主要来源,提出了一种对虚拟仪器进行非线性校正的支持向量机方法。该方法能够克服神经网络处理小样本问题的不足,具有较高的泛化能力。实验表明,用支持向量机算法解决虚拟仪器非线性问题是有效的。

With the wide use of virtual instruments, the error control problem is becoming more and more prominent. The traditional nonlinear calibration method of virtual instrument is based on artificial neural network. Due to the inherent faults of ANN, its application is affected. Support vector machine is a newly developed machine learning algorithm which is applied in lots of fields. This paper analyzes the main nonlinear error sources of virtual instrument, and presents a nonlinear calibration method with Support Vector Machine. This method can solve the problem caused by small sampling in artificial neural network, and has better performance of generalization. The experiment result shows that SVM algorithm is effective to nonlinear calibration of virtual instrument.