[1]张冬青,刘欢,张云清.基于Q-RBF神经网络模型的国产大豆价格预测研究[J].大豆科学,2017,36(01):143-149.[doi:10.11861/j.issn.1000-9841.2017.01.0143]
　ZHANG Dong-qing,LIU Huan,ZHANG Yun-qing.Forecasting Chinese Domestic Soybean Price Based on Q-RBF Neural Network Model[J].Soybean Science,2017,36(01):143-149.[doi:10.11861/j.issn.1000-9841.2017.01.0143]
点击复制

基于Q-RBF神经网络模型的国产大豆价格预测研究

《大豆科学》[ISSN:1000-9841/CN:23-1227/S] 卷: 第36卷 期数: 2017年01期 页码: 143-149 栏目: 出版日期: 2017-01-20
Title:
Forecasting Chinese Domestic Soybean Price Based on Q-RBF Neural Network Model
作者:
张冬青; 刘欢; 张云清
（南京农业大学 工学院，江苏 南京 210031）
Author(s):
ZHANG Dong-qing; LIU Huan; ZHANG Yun-qing
（College of Engineering, Nanjing Agricultural University, Nanjing 210031, China）
关键词:
预测; Q-RBF神经网络; 梯度下降法; 遗传算法; 概率密度函数
Keywords:
Forecasting;  Quantile regression-radial basis function (Q-RBF) neural network;  Gradient descent method;  Genetic algorithm;  Probability density function
DOI:
10.11861/j.issn.1000-9841.2017.01.0143
文献标志码:
A
摘要:
大豆是重要的经济作物，同时也是我国市场化和国际化程度最高的大宗农产品，对其价格进行预测具有重要意义。采用Q-RBF神经网络模型对国产大豆价格进行预测，该模型具有如下两个特点：（1）通过分位数回归功能来描述大豆在不同价格水平下的分布特征；（2）通过RBF神经网络结构来刻画大豆价格的非线性关系。在模型参数优化时，由于遗传算法是一种全局搜索优化方法，但是搜索速度慢、对初始值具有一定依赖性；而梯度下降法具有收敛快，对初始值没有特定要求等优点，所以本文提出遗传算法与梯度下降法相结合的混合改进算法，其基本思想是利用梯度下降法的局部寻优能力加快遗传算法的收敛速度。采用2010年1月-2015年12月的国产大豆月度价格数据进行预测研究，结果表明，算法收敛速度较快，模型预测精度较高，是可以泛化应用的预测模型。
Abstract:
Soybean is an important cash crop, and it is also the most important agricultural product with the highest degree of marketization and internationalization character in China. So it is essential to forecast the soybean price. A Quantile-RBF (Q-RBF) neural network model is proposed to predict Chinese domestic soybean price in this paper. The model has two characteristics as follows: (1) Quantile regression models describe the distribution over the range of the soybean price; (2) RBF neural networks approximate the nonlinear part of soybean price. The parameters of Q-RBF neural network model can be optimized through the genetic algorithm (GA) and the gradient descent method. GA is a global optimization method, however, it might be slow in convergence. On the contrary, the gradient descent method quickly converges to an optimal solution, but may converge to a local minimum or maximum and is not efficient in discontinuous problems. Therefore, an improved algorithm combining GA with gradient descent method is proposed in this paper. In the improved algorithm, the gradient descent method is used to improve the convergence efficiency of GA. The data of monthly soybean price from Jan. 2010 to Dec. 2015 were analyzed. Experimental results demonstrated that the Q-RBF neural network model and improved algorithm were accurate and effective.

参考文献/References:

［1］奚晓菁. 我国大豆价格波动影响因素分析［D］. 昆明: 云南财经大学, 2013. (Xi X J. Analysis on the influencing factors of soybean price fluctuation in China［D］. Kunming: Yunnan University of Finance and Economics, 2013.)

［2］张婷. 基于ARIMA模型的国际粮食短期价格分析预测—以大豆为例［J］. 价格月刊, 2016, (470): 28-32. (Zhang T. Analysis and prediction of short term international grain price based on ARIMA model: A case study of soybean［J］. Prices Monthly, 2016, (470): 28-32.)
［3］朱婧, 范亚东, 徐勇. 基于改进GM(1,1)模型的中国大豆价格预测［J］. 大豆科学, 2016, 35(2): 315-319. (Zhu J, Fan Y D, Xu Y. Soybean price prediction in china based on modified GM(1,1) model［J］. Soybean Science, 2016, 35(2): 315-319.)
［4］程文晓. 我国大豆期货价格的预测分析［D］. 兰州: 兰州大学, 2014. (Cheng W X. Forecasting of soybean futures prices in China［D］. Lanzhou: Lanzhou University, 2014.)
［5］石波, 张冬青, 马开平，等. 改进RBF神经网络在我国大豆价格预测中的应用研究［J］. 大豆科学, 2016, 35(2): 310-314. (Shi B, Zhang D Q, Ma K P, et al. Soybean price prediction in China based on improved RBF neural network［J］. Soybean Science, 2016, 35(2): 310-314.)
［6］毛学峰, 贾伟. 大豆及制成品动态特征价格的实证研究［J］. 农业技术经济, 2016(7):73-80. (Mao X F, Jia W. An empirical study on the price of dynamic characteristics of soybean and manufactured goods［J］. Journal of Agrotechnical Economics, 2016(7):73-80.)
［7］Berwald D, Havenner A. Evaluating state space forecasts of soybean complex prices ［M］// Aoki M, et al .Applications of computer aided time series modeling. New York:Springer-Verlag, Inc. 1997.
［8］Arnade C, Hoffman L.The impact of price variability on cash/futures market relationships: Implications for market efficiency and price discovery［J］. Journal of Agricultural & Applied Economics, 2015, 47(4): 539-559.
［9］Adrangi B, Chatrath A, Raffiee K. Price discovery in the soybean futures market［J］. Journal of Business & Economics Research, 2011, 4(6): 77-88.
［10］Koenker R, Bassett G W. Regression quantiles［J］. Econometric, 1978, 46:33-50.
［11］Taylor J W. A quantile regression neural network approach to estimating the conditional density of multiperiod returns［J］. Journal of Forecasting, 2000, 19(4): 299-311.
［12］许启发, 蒋翠侠. 分位数局部调整模型及应用［J］. 数量经济技术经济研究, 2011(8): 115-133. (Xu Q F, Jiang C X. Quantile partial adjustment model and its application［J］. The Journal of Quantitative & Technical Economics, 2011(8): 115-133.)
［13］Cannon A J. Quantile regression neural networks: Implementation in R and application to precipitation downscaling［J］. Computers & Geosciences, 2011, 37: 1277-1284.
［14］何耀耀, 许启发, 杨善林等. 基于RBF神经网络分位数回归的电力负荷概率密度预测方法［J］. 中国电机工程学报, 2013, 33(1): 93-98. (He Y Y, Xu Q F, Yang S L, et al. A power load probability density forecasting method based on RBF neural network quantile regression［J］. Proceedings of the CSEE, 2013, 33(1): 93-98.)
［15］Mok T K, Liu H M, Ni Y X, et al. Tuning the fuzzy damping controller for UPFC through genetic algorithm with comparison to the gradient descent training［J］.Electrical Power and Energy Systems, 2005, 27: 275-283.
［16］刘欢, 张冬青. 基于分位数回归的国产大豆价格影响因素分析［J］. 大豆科学, 2014, 33(5): 759-763. (Liu H, Zhang D Q. Analysis on influencing factors of domestic soybean price based on quantile regression［J］. Soybean Science, 2014, 33(5): 759-763.)

相似文献/References:

[1]祁雪莲,邓华玲,徐 丹,等.数量化理论在预测大豆食心虫发生概率上的应用[J].大豆科学,2012,31(04):640.[doi:10.3969/j.issn.1000-9841.2012.04.026]
　QI Xue-lian,DENG Hua-ling,XU Dan,et al.Application of Quantification Theory in Predicting Occurrence Probability of Soybean Pod Borer[J].Soybean Science,2012,31(01):640.[doi:10.3969/j.issn.1000-9841.2012.04.026]
[2]姚鑫淼 张瑞英 李霞辉 程爱华 张晓波 　 高春霞.近红外透射光谱法(NITS)分析大豆品质的研究*[J].大豆科学,2006,25(04):417.[doi:10.11861/j.issn.1000-9841.2006.04.0417]
　Yao Xinmiao Zhang Ruiying Li Xiahui Cheng Aihua Zhang Xiaobo Gao Chunxia.STUDY ON ANALYSIS OF SOYBEAN QUALITY BY NEARINFRARED TRANSMITTANCE SPECTROSCOPY[J].Soybean Science,2006,25(01):417.[doi:10.11861/j.issn.1000-9841.2006.04.0417]
[3]刘志芳.大豆产量预测预控数学模型的研究[J].大豆科学,1991,10(03):187.[doi:10.11861/j.issn.1000-9841.1991.03.0187]
　[J].Soybean Science,1991,10(01):187.[doi:10.11861/j.issn.1000-9841.1991.03.0187]
[4]程志明,矫洪双,陆修德,等.大豆菌核病流行预测研究[J].大豆科学,1994,13(01):48.[doi:10.11861/j.issn.1000-9841.1994.01.0048]
　[J].Soybean Science,1994,13(01):48.[doi:10.11861/j.issn.1000-9841.1994.01.0048]
[5]徐鑫洲,马开平.基于系统动力学的我国大豆价格预测分析[J].大豆科学,2018,37(05):787.[doi:10.11861/j.issn.1000-9841.2018.05.0787]
　XU Xin-zhou,MA Kai-ping.Prediction and Analysis of Soybean Price in China Based on System Dynamics[J].Soybean Science,2018,37(01):787.[doi:10.11861/j.issn.1000-9841.2018.05.0787]
[6]马宏伟,白荻,李静,等.中国大豆2021—2025年消费量和生产量预测分析[J].大豆科学,2022,41(03):358.[doi:10.11861/j.issn.1000-9841.2022.03.0358]
　MA Hong-wei,BAI Di,LI Jing,et al.Prediction and Analysis of China’s Soybean Consumption and Production in 2021—2025[J].Soybean Science,2022,41(01):358.[doi:10.11861/j.issn.1000-9841.2022.03.0358]

备注/Memo

基金项目：国家自然科学基金（71301077）；南京农业大学中央高校基本科研业务费人文社会科学基金（SK2014011）。

第一作者简介：张冬青(1971-)，女，博士，副教授，主要从事预测与决策、农业系统工程研究。E-mail:hollycase@163.com。

更新日期/Last Update: 2017-03-15