计算机与现代化

• 算法设计与分析 • 上一篇    下一篇

基于时间序列神经网络的鲜切花价格指数短期预测

  

  1. (云南农业大学云南省高校农业信息技术重点实验室,云南昆明650201)
  • 收稿日期:2018-05-03 出版日期:2019-05-14 发布日期:2019-05-14
  • 作者简介:彭伟(1990-),男,四川资中人,硕士研究生,研究方向:计算机应用,神经网络,大数据分析,E-mail: 915643695@qq.com。
  • 基金资助:
    国家农村信息化示范省省级综合信息资源中心建设项目(2014AB017)

Short-term Prediction of Fresh Cut Flower Price Index Based on Time Series Neural Network

  1. (University Key Laboratory of Agricultural Information Technology in Yunnan, Yunnan Agricultual University, Kunming 650201, China)
  • Received:2018-05-03 Online:2019-05-14 Published:2019-05-14

摘要: 鲜切花价格指数是反映鲜切花市场现状的风向标,研究鲜切花价格指数变化,掌握鲜花市场的动态和规律性具有重要意义。本文针对具有时序特点的鲜切花价格指数,基于BP模型中的L-M优化算法构建鲜切花价格指数短期预测模型,采用tansig和purelin作为各层之间的传递函数,利用时间序列分析方法确定输入层的神经元个数,通过实验数据对比来确定隐含层的神经元个数。采用平均绝对误差、平均相对误差和均方根误差这3个评价指标对模型的预测精度进行检验,实验结果表明所构建模型是有效的和具有实际应用价值的。

关键词: 鲜切花价格指数, BP模型, 短期预测, 评价指标

Abstract:  The fresh cut flower price index is a trend indicator reflecting the current status of the fresh cut flower market. It is of great significance to study the change of fresh cut flower price index and grasp the dynamics and regularity of flower market. Aiming at the sequence characteristics of fresh cut flower price index, this paper constructs the cut flower price index short-term forecasting model based on the L-M optimization algorithm of BP model. The model uses tansig and purelin as the transfer function between layers, uses the time series analysis method to determine the number of input layer neurons, and by comparison with experimental data to determine the number of hidden layer neurons. Three evaluation indexes of mean absolute error, mean relative error and root mean square error are used to test the prediction accuracy of the model. The experimental results show that the model is effective and has practical application value.

Key words:  fresh cut flower price index, BP model, short-term prediction, evaluation indexes

中图分类号: