环氧树脂灌封胶的性能预测及配方优化

Prediction of Properties and Optimization of Formula for Epoxy Pouring Sealant

支持向量机(SVM)是一种新型的机器学习方法,以结构风险最小化原则取代传统机器学习方法中的经验风险最小化原则,在小样本的机器学习中显示出了优异的性能。将SVM应用于双组分环氧树脂灌封胶的研制。通过对双组分环氧树脂灌封胶配方的学习,建立SVM推理模型,并结合穷举法对配方进行优化,结果表明所建SVM推理模型具有一定的预测能力,展示了其优越性和推广前景,可应用于胶粘剂配方的研制,对配方优化起到一定的指导作用。

Support Vector Machines （SVM） is a sort of novel learning method, which is based on Structural Risk Minimization principle instead of traditional machine learning based on Empirical Risk Minimization principle. SVM exhibited the excellent properties in learning limited samples. In the paper, SVM was applied to study the preparation of Dual - epoxy pouring sealant. The formula of Dual - epoxy pouring sealant was learned and the SVM model was founded, in the mean time, formula for epoxy pouring sealant was optimized by SVM and exhaustive method. The results demonstrated the effectiveness of using the support vector machine in pouring sealant properties forecast. It also pointed out that SVM was a promising learning method, which could be used to forecast the formula of adhesive and provide an important basis for opti- mization of formula.