首页|基于主成分分析的果蝇算法优化支持向量机回归的红枣产量预测

基于主成分分析的果蝇算法优化支持向量机回归的红枣产量预测

扫码查看
随着大数据技术和人工智能的快速发展,针对当前红枣产量预测模型精度低、模型优化时间过长等问题,以山西省1993-2020 年的红枣产量及17 个维度的因素作为基础数据,提出一种基于主成分分析的果蝇算法优化支持向量机回归(princi-pal component analysis-fruit fly optimization algorithm-support vector regression,PCA-FOA-SVR)的红枣产量预测模型.首先利用主成分分析(principal component analysis,PCA)对数据进行降维处理,以5 维的指标作为输入变量,产量作为输出变量;其次以支持向量机回归(support vector regression,SVR)为基础模型,利用果蝇优化算法(fruit fly optimization algorithm,FOA)对SVR参数惩罚因子c和核函数参数g进行寻优,构建PCA-FOA-SVR模型.对试验结果进行验证.发现PCA-FOA-SVR的均方根误差(root mean square error,RMSE)、平均绝对误差(mean absolute error,MAE)、决定系数R2分别为3.11、3.01、0.96,SVR的各指标分别为5.33、4.07、0.9,分别提高了41.7%、26%、6.7%,最后通过GM(1,1)对各维度的数据进行预测,利用PCA-FOA-SVR模型对未来10 年山西省红枣产量进行预测,结果显示在2025 年红枣产量会达到一个峰值,对后续相关研究提供了一定的科学依据.
Fruit Fly Algorithm Optimized Support Vector Regression for Yield Prediction of Jujube Based on Principal Component Analysis
With the rapid development of Big data technology and artificial intelligence,in view of the problems such as low accura-cy of the current jujube yield prediction model and long optimization time of the model,the jujube yield in Shanxi Province from 1993 to 2020 and the factors in 17 dimensions were taken as the basic data,a fruit fly algorithm optimized support vector regression(PCA-FOA-SVR)model was proposed for predicting jujube yield based on principal component analysis.Firstly,principal component analy-sis(PCA)was used to reduce the dimensionality of the data,with 5-dimensional indicators as input variables and production as output variables.Secondly,based on the support vector machine regression model(SVR),the fruit fly optimization algorithm(FOA)was used to optimize the SVR parameter penalty factor c and kernel function parameter g,and a PCA-FOA-SVR model was constructed.The test results were verified.It is found that the root-mean-square deviation(RMSE),mean absolute error(MAE),and coefficient of deter-mination R2 of PCA-FOA-SVR are 3.11,3.01,0.96,respectively,and the indicators of SVR are 5.33,4.07,0.9,increased by 41.7%,26%,and 6.7%,respectively.Finally,GM(1,1)was used to predict the data of various dimensions,and the PCA-FOA-SVR model was used to predict the jujube production in Shanxi Province in the next 10 years.The results show that the jujube produc-tion will reach a peak in 2025,which provides a certain scientific basis for subsequent related research.

yield prediction of red datessupport vector machine regression(SVR)fruit fly optimization algorithm(FOA)princi-pal component analysis(PCA)

李晋泽、赵素娟、李宁、李俊成、刘森、马继东

展开 >

东北林业大学机电工程学院,哈尔滨 150040

太原理工大学矿业工程学院,太原 030024

红枣产量预测 支持向量机回归(SVR) 果蝇算法(FOA) 主成分分析(PCA)

国家自然科学基金

31870537

2024

科学技术与工程
中国技术经济学会

科学技术与工程

CSTPCD北大核心
影响因子:0.338
ISSN:1671-1815
年,卷(期):2024.24(4)
  • 19