Abstract: This paper presents an improved algorithm about Support Vector Machine(SVM) and a pattern recognition algorithm based on distance to solve the problem of the use prediction of equipment.It finds the optimal hyperplane of SVM using the training data,and uses a characteristic set of test data as classification criteria to predict the classification results.The combination of the classification results and the probability is used as the input of pattern recognition algorithms,and the output is a pattern.Experimental results show that,compared with traditional algorithms,the accuracy of pattern recognition algorithm using improved SVM classification as input is better than traditional ones,it can be widely used in pattern recognition algorithms with binary input.

Key words: Support Vector Machine(SVM), cosine vector, classifier, fuzzy feature, prediction model

摘要： 为解决设备使用预测的问题,给出支持向量机(SVM)的改进算法及基于距离的模式识别算法。使用训练数据得到SVM的最优分类超平面,运用确认数据的特征集作为分类标准预测分类结果,将分类结果与概率相结合作为模式识别算法的输入,算法输出为某个固定模式。实验结果表明,与传统算法相比,以改进的SVM分类结果为输入的模式识别算法准确性更高,可广泛应用在二值输入的模式识别算法中。

关键词: 支持向量机, 余弦向量, 分类器, 模糊特征, 预测模型

CLC Number: 

• TP391