摘要
针对气候舱法检测人造板家具甲醛释放浓度,存在实验周期长,耗能高的问题。对此提出互补集合经验模态分解(Complementary Ensemble Empirical Mode Decomposition,CEEMD)与支持向量机(Support Vector Machine,SVM)相结合的多步预测模型,对舱内家具的甲醛释放浓度进行预测。先利用CEEMD对甲醛浓度数据进行降噪,再采用SVM建立预测模型,并使用多种群遗传算法(Multiple Population Genetic Algorithm,MPGA)优化SVM参数。与传统SVM及一阶衰减模型对比结果表明,所建CEEMD-MPGA-SVM预测模型具有更高的预测精度。
Aiming at the problems of long experimental period and high energy consumption in the detection of formaldehyde emission concentration from wood-based furniture by the environment chamber,a multi-step prediction model combining complementary ensemble empirical mode decomposition(CEEMD)and support vector machine(SVM)is proposed to predict the formaldehyde concentration released from furniture in the environment chamber.The formaldehyde concentration data are firstly noise-reduced using CEEMD,then the prediction model is built by SVM,and the SVM parameters are optimized by using multiple population genetic algorithm(MPGA).Compared with the traditional SVM and first-order decay model,the proposed CEEMD-MPGA-SVM prediction model has higher prediction accuracy.
作者
孙钦乾
刘萌萌
郑焕祺
周玉成
Sun Qinqian;Liu Mengmeng;Zheng Huanqi;Zhou Yucheng(Shandong Jianzhu University,Jinan,Shandong 250101,China;Shandong Institute for Product Quality Inspection)
出处
《计算机时代》
2023年第1期40-43,共4页
Computer Era
基金
《家具中挥发有机物的筛查检测方法气相色谱质谱法》国家标准制定计划项目(20203844-T-607)
山东建筑大学博士基金(X21110Z)。
