基于混合特征LLE融合与SVM的质量异常模式识别

针对生产过程中存在的异常模式识别的问题,提出基于LLE融合与支持向量机的质量异常模式识别方法。首先从动态数据流中提取其原始特征、统计特征、几何特征并将其进行混合,形成动态数据流的混合特征,然后利用LLE算法对混合特征进行降维,将降维后的特征集作为MSVM分类器的输入进行训练,同时采用粒子群算法对MSVM分类器进行参数寻优。最后用训练好的模型对动态数据流进行异常模式的识别。并将所提方法与单一类型特征方法、混合特征方法的识别模型进行比较,仿真结果和应用实例表明,所提方法的识别精度较高,可用于生产过程的质量异常模式识别中。

Quality Abnormal Pattern Recognition Based on Hybrid Feature LLE Fusion and SVM

Aiming at the problem of abnormal pattern recognition in production process,a method of quality abnormal pattern recognition based on LLE fusion and support vector machine is proposed.Firstly,the original feature,statistical feature and geometric feature are extracted from the dynamic data stream and mixed to form the mixed feature of the dynamic data stream.Then,the LLE algorithm is used to reduce the dimension of the mixed feature.The reduced feature set is trained as the input of the MSVM classifier.At the same time,the particle swarm optimization algorithm is used to optimize the parameters of the MSVM classifier.Finally,the trained model is used to recognize abnormal patterns in dynamic data streams.The proposed method is compared with single type feature method and mixed feature method.The simulation results show that the proposed method has high recognition accuracy and can be used in quality anomaly pattern recognition of production process.