基于最小二乘支持向量机的北京市肉类需求量预测研究

发布时间：2018-10-31 16:00

【摘要】：作为拥有1921万常住人口的大型城市,北京市每年的肉类需求量巨大。2010年,北京市肉类需求量为547086吨,其中对猪牛羊肉的需求量分别为388542吨、88080吨和70464吨。肉类及其制品作为人们日常生活不可或缺的食品,对于保障民生具有重要作用。为了保证肉类及其制品的供应,减少资源浪费,有必要对其需求量进行有效预测。本文研究了肉类需求量预测模型的选择、关键影响因素的确定,并应用所选预测模型对北京市肉类需求量进行了预测。主要内容如下：首先,对比国内外常用的需求预测方法。传统预测方法：回归预测方法、时间序列预测方法,以及智能化方法：人工神经网络预测方法、最小二乘支持向量机预测方法,进行分析,并比较优缺点,选择最小二乘支持向量机预测模型进行北京市肉类需求量的预测。其次,选取影响北京市肉类需求量的关键因素。分析北京市肉类需求量的变化趋势以及肉类食品的消费结构,确定主要预测目标。总结归纳所有影响北京市肉类需求量的因素,根据相关性分析和数据可得性,选取人口数、每人每年现金收入和肉类食品零售价格指数为关键因素。最后,应用最小二乘支持向量机预测模型对北京市肉类需求量进行预测。将所得结果与真实值、多元回归方法预测值和BP神经网络模型预测值进行比较,证实此模型具有较高的预测精度。并应用最小二乘支持向量机预测北京市2015年、2020年的肉类需求量。结果表明,猪肉需求量在未来十年会有较大增幅,牛羊肉需求量则有下降趋势。论文的研究工作可以为北京市的肉类食品企业发展规划、北京市冷链物流系统规划提供较为准确地肉类需求量数据,为肉类需求量预测模型的选择和计算的进一步研究,提供了参考和借鉴。
[Abstract]:As a large city with a permanent population of 19.21 million, Beijing has a huge demand for meat every year. In 2010, the demand for meat in Beijing was 547086 tons, including 388542 tons for pigs, 88080 tons for beef and 70464 tons for beef and mutton. As an indispensable food in people's daily life, meat and its products play an important role in protecting people's livelihood. In order to ensure the supply of meat and its products and reduce the waste of resources, it is necessary to forecast the demand of meat and its products effectively. In this paper, the selection of forecasting model of meat demand and the determination of key influencing factors are studied, and the forecast model is used to predict the demand of meat in Beijing. The main contents are as follows: first of all, the common demand forecasting methods at home and abroad are compared. Traditional prediction methods: regression forecasting, time series prediction, and intelligent methods: artificial neural network prediction method, least squares support vector machine prediction method, are analyzed, and the advantages and disadvantages are compared. The least square support vector machine (LS-SVM) model is used to predict the meat demand in Beijing. Secondly, select the key factors that affect the meat demand in Beijing. The change trend of meat demand and the consumption structure of meat food in Beijing are analyzed, and the main forecast target is determined. All the factors affecting meat demand in Beijing are summarized. According to the correlation analysis and the availability of data, the population size, the annual cash income per person and the retail price index of meat food are selected as the key factors. Finally, the least square support vector machine (LS-SVM) model is used to predict the meat demand in Beijing. The results obtained are compared with the real values, the multivariate regression method and the BP neural network model. It is proved that this model has a high prediction accuracy. And the least square support vector machine is used to forecast the meat demand in Beijing in 2015 and 2020. The results showed that the demand for pork would increase greatly in the next ten years, and the demand for beef and mutton would decrease. The research work in this paper can provide more accurate data of meat demand for the development planning of meat food enterprises in Beijing and cold chain logistics system planning in Beijing, and further research on the selection and calculation of meat demand forecasting models. Reference and reference are provided.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2013
【分类号】：F426.82;F224

【参考文献】