Prediction of air pollution index (API) using support vector machine (SVM)
Prediction of air pollution index (API) using support vector machine (SVM)
复制标题
DOI：
10.1016/j.jece.2019.103208
复制
发表时间：
2020-06-01
影响因子：
7.7
通讯作者：
Ahmad, Z.
中科院分区：
文献类型：
--
作者：
Leong, W. C.;Kelani, R. O.;Ahmad, Z.
现有的大气污染指数计算方法复杂且耗时。因此，需要提出新的准确、高效的建模技术。因此，本研究提出一种支持向量机对空气污染指数进行建模。影响支持向量机模型性能的主要参数有三个：惩罚因子(C)、正则化参数（epsilon）和所使用的核函数类型。然而，本研究仅对核函数模型参数进行了研究。采用平方和误差（SSE）、平方和误差均值（MSSE）和决定系数（R-2）对模型结果进行分析。结果表明，采用径向基函数（RBF）核函数建立的模型能够有效、准确地解决复杂空气污染指数的和方误差（SSE）、均方和误差（MSSE）和决定系数（R-2）分别为2008、3.1.4440和0.9843的建模问题。
The existing methods of calculating air pollution index are complex and time consuming. Therefore new accurate and efficient modeling techniques need to be proposed. Thus, a support vector machine is proposed in this study to model the air pollution index. There are three main parameters affecting the performance of the support vector machine model: penalty factor (C), regularization parameter (epsilon) and the type of kernel function used. However, in this study, only kernel functions model parameters are investigated. The results of the model are analyzed by using sum of squares error (SSE), mean of sum of squares error (MSSE) and coefficient of determination (R-2). It is found that the proposed model using radial basis function (RBF) kernel function effectively and accurately able to solve the problem of complex air pollution index modeling with sum square error (SSE), mean sum square error (MSSE) and coefficient of determination (R-2) of 2008, 3.1.4440 and 0.9843 respectively.