基于智能算法鲜切花价格预测平台设计的探索与实践

Exploration and Practice of Price Forecasting Platform for Fresh Cut Flowers Based on Intelligent Algorithm

针对市场上对各类鲜切花零售价格、销售量和销售率等数据收集不齐全的问题,拟设计一个零售商和买家参与的基于智能算法鲜切花价格预测平台:通过零售商和买家交易来收集鲜切花等数据,作为数据输入源进行基于智能算法的鲜切花价格预测,预测的结果反馈给鲜切花种植户、种植企业保障其经济效益,同时为研究人员深入研究提供数据。该研究主要对鲜切花预测的关键技术进行探究,根据输入层数据的规模,分别使用径向基函数神经网络(RBF)、广义回归神经网络算法(GRNN)构建的鲜切花价格预测模型,以玫瑰鲜切花作为研究对象,并使用斗南花卉市场公布的部分数据作为模型的数据输入源进行模拟预测。结果表明:径向基函数神经网络(RBF)、广义回归神经网络算法(GRNN)分别适用于不同级别规模的输入层数据,预测率保持在85%~95%。

In order to solve incomplete collecting data of all kinds of fresh cut flowers,such as retail price,sales volume,sales rate and so on in the market,based on intelligent algorithm,aprice prediction platform of fresh cut flowers is designed for retailers and buyers：the data of fresh cut flowers are collected via the trading of retailers and buyers,then using these data as input sources to predict the price of fresh cut flowers based on intelligent algorithm,at last,the forecast results not only are fed back to fresh cut flower growers and plant enterprises to ensure their economic benefits,but also can provide in-depth research data for researchers.In this study,the key techniques of fresh cut flower prediction were explored.According to the size of the input layer data,Radical Basic Function（RBF）,General Regression Neural Network（GRNN）were applied respectively to build the cut flower price forecasting model.Taking the rose cut flowers as example,with the data published by Dounan Flower Market as input source of the two prediction models.The experimental results showed that RBF and GRNN were suitable for different levels scale of input data sets and the forecast accuracy rate remained at 85%-95%.