RVM Soft Sensing Model Based on Optimized Combined Kernel Function

doi:10.16182/j.issn1004731x.joss.201801035

Abstract

Abstract: An RVM spft sensingmodeling method based onthe optimizedcombined kernel functionis proposed.In order to simultaneously get better prediction and sparsity, a fitness function synthesizing regression accuracy and sparsity is created while constructing a combined kernel functionfor RVM.The genetic algorithm is used to optimize the weights and kernel parametersof the RVMcombined kernel.The proposed method is used totomodela cleavage-recovery unit in the production process of Bisphenol-A.The results show that it can guarantee better sparsity andregression accuracy than the general SVM combinedkernel model andGA-RVM single kernel model.

Key words: bisphenol A, relevance vector machine, genetic algorithms, fitness function, kernel parameters, mixed kernel function

CLC Number:

TP274

Zhang Yanan, Yang Huizhong. RVM Soft Sensing Model Based on Optimized Combined Kernel Function[J]. Journal of System Simulation, 2018, 30(1): 272-277.

References

[1] Petr Kadleca.Data-driven Soft Sensors in the process industry[J]. Computers and Chemical Engineering (S0098-1354), 2009(33): 795-814.
[2] Sheng-weiFei. Kurtosis forecasting of bearing vibration signal based on the hybrid model of empirical mode decomposition and RVM with artificial bee colony algorithm[J]. Expert Systems with Applications (S0957-4174), 2015(42): 5011-5018.
[3] YILMAZ I.Comparison of landslide susceptibility mapping methodologies for Koyulhisar, Turkey: conditional probability, logistic regression, artificial neural networks, and support vector machine[J]. Environmental Earth Sciences(S1866-6280), 2010, 61(4): 821-836.
[4] Damoulas T, Ying Y, Girolami M A, et al.Inferring sparse kernel combinations and relevance vectors: an application to subcellular localization of proteins[C]// Proceedings of the 7th International Conference on Machine Learning and Applications (ICMLA2008). San Diego, USA: IEEE, 2008: 577-582.
[5] 白霜. 基于PSO优化混合RVM模型的进出口贸易预测算法[J]. 计算机与现代化, 2014(8): 110-118.Shuang Bai. Import and Export Trade Prediction Algorithm Based on Multi-Kernel Weighting RVM Model Optimized by PSO[J]. Computer and Modernization, 2014(8): 110-118.
[6] 肖汉杰, 桑秀丽. 相关向量机超参数优化的网络安全态势预测[J]. 计算机应用, 2015, 35(7): 1888-1891.Xiao Hanjie, Sang Xiuli. Network Security Situation Prediction based on Hyper Parameter Optimization of Relevance Vector Machine[J]. Journal of Computer Applications, 2015, 35(7): 1888-1891.
[7] 李刚, 王贵龙, 薛惠锋. RVM核参数的遗传算法优化方法[J]. 控制工程, 2010, 17(3): 335-338.Gang Li, Guilong Wang, HuifengXue. GA Optimizing Method to Kernel Function Parameters of RVM[J]. Control Engineering, 2010, 17(3): 335-338.
[8] Tipping E.Sparse Bayesian Learning and the Relevance Vector Machine[J]. Journal of Machine Learning Research(S1532-4435), 2001(1): 211-244.
[9] Qianqian Zhao, Kewen Xia, Zhaozheng Hu, et al.Relevance Vector Machine and Its Application in Gas Layer Classification[J]. Journal of Computational Information Systems (S1553-9105), 2013, 9(20): 8343-8350.
[10] 赵榈. 相关向量机优化方法的研究[D]. 南宁: 广西大学, 2012.Zhao Lü. Optimization Methodology Research of Relevance Vector Machine[D]. Nanning: Guangxi University, 2012.
[11] Chao Hu, Gaurav Jain, Craig Schmidt, et al.Online estimation of lithium-ion battery capacity using sparse Bayesian learning[J]. Journal of Power Sources (S0378-7753), 2015(289): 105-133.
[12] 陈果, 邓堰. 遗传算法特征选取中的几种适应度函数构造新方法及其应用[J]. 机械科学与技术, 2012, 30(1): 124-132.Guo Chen, Yan Deng. Several New Methods for Features Extraction Based omn Genetic Algorithm and Their Application [J]. Mechanical Science and Technology for Aerospace Engineering, 2012, 30(1): 124-132.
[13] 朱永利, 尹金良. 组合核相关向量机在电力变压器故障诊断中的应用研究[J]. 中国电机工程学报, 2013, 22(33): 68-74.Zhu Yongli, Yin Jinliang. Study on Application od Multi-kernel Learning Relevance Vector Machines in Fault Diagnosis of Power Transformers[J]. Proceedings of the CSEE, 2013, 22(33): 68-74.