恶臭气相色谱信号的分类算法研究

Research on classification algorithm for malodorous gas chromatography signals

工业化的快速发展带来了恶臭污染问题,而现有的恶臭气相色谱仪(gas chromatography, GC)面临着因算力分配不均而导致的效率低下的问题。通过对恶臭气相色谱信号进行分类并分派给为每种峰形定制的解析算法可以使算力分配得以优化。文中致力于恶臭气相色谱信号分类算法的研究,着重进行了恶臭气相色谱信号的处理,并对比了K最近邻(K-nearest neighbor, KNN)、支持向量机(support vector machines, SVM)、随机森林(random forest, RF)3种分类算法的性能。首先对信号进行预处理,提取了特征点并计算特征参数,以特征参数为分类依据代入算法程序并检验算法的准确率,以此来评价每种算法的性能。实验结果显示RF算法拥有较好的应用效果和发展潜力。该文可为实现算力的均衡分配提供一些参考依据。

The rapid development of industrialization has brought about the problem of odor pollution,and current malodorous gas chromatography face the problem of low efficiency due to unbalanced distribution of computing power.Computational power allocation can be optimized by classifying malodorous gas chromatography signals and assigning them to resolution algorithms customized for each peak shape.This paper is devoted to the research of malodorous gas chromatography signals classification algorithm,focusing on the processing of malodorous gas chromatography signals,and compared the performance of"K-nearest neighbor(KNN)","support vector machines(SVM)","random forest(RF)"classification algorithms.Firstly,the signal is preprocessed,the feature points are extracted and the characteristic parameters are calculated,and the characteristic parameters are used as the classification basis to be substituted into the algorithm program and the accuracy of the algorithm is checked to evaluate the performance of each algorithm.The experimental results show that the RF algorithm has good application effect and development potential.This paper can provide a reference for realizing the balanced distribution of computing power.