摘要
基于福建南平、三明和龙岩三大烟区代表性样点烤烟产值及相关调查资料,借助神经网络和双重筛选逐步回归2种数学模型筛选福建烟区主要烤烟品种K326、云烟87(85)的产值估测模型;并借助GIS与烤烟产值最佳估测模型集成技术,开展福建烟区烤烟产值分区.结果表明:神经网络模型对品种K326、云烟87(85)产值的模拟精度分别高达99.74%和99.72%,表明该模型可作为福建烟区烤烟产值较理想的估测模型;福建烟区烤烟K326的高产值和低产值区面积分别占全省烟区烤烟生态适宜用地总面积的26.36%和54.43%;云烟87(85)高产值和低产值区面积分别占全省烟区烤烟生态适宜用地总面积的48.77%和30.25%,表明福建烟区云烟87(85)高产值区面积明显大于K326.
Based on the output value and some survey data of flue-cured tobacco of Nanping,Sanming and Longy an in Fujian,using neural network and dual screening stepwise regression model,the ideal output value estimation models of flue-cured tobacco such as K326 and Yunyan 87 (85) in Fujian were screened.With the aid of GIS and integration technology of the best estimation model of flue-cured tobacco output value,the division of flue-cured tobacco output values in Fujian was carried on.The results were as follows:the estimation accuracy of neural network were 99.74% and 99.72%,respectively; the highest output value region and the low output value region of K326 accounted for 26.36% and 54.43% of the total area of Fujian flue-cure tobacco ecological suitable land,respectively; the highest output value region and the low output value region of Yunyan 87(85) accounted for 48.77% and 30.25% of the total area of Fujian flue-cure tobacco ecological suitable land,respectively.The results indicated that the highest out-put value region of Yunyan 87 (85) was significantly larger than that of K326.
出处
《福建农林大学学报(自然科学版)》
CSCD
北大核心
2014年第6期561-567,共7页
Journal of Fujian Agriculture and Forestry University:Natural Science Edition
基金
福建省科技厅重大专项资助项目(2012NZ0002)
关键词
福建
烤烟
产值
估测模型
分区
Fujian
flue-cured tobacco
output value
estimating model
division