摘要
为了实现对蓄水坑灌苹果园根区土壤水分的定量监测,基于最小二乘向量机模型、BP神经网络模型和增量逆传播学习算法优化的BP神经网络模型,建立了以土壤初始贮水量、灌水后时段长度、时段内灌水量、时段内降雨量和蒸发蒸腾量为输入项,以蓄水坑灌果园根区贮水量为输出项的LSSVM、BP和IBP-BP模型,对田间土壤贮水量进行预测,并采用田间实测数据对模型进行率定和验证。结果表明:LSSVM、IBP-BP和BP模型的平均相对误差分别为6.53%、3.64%和5.98%。IBP-BP模型的预测精度最高,建议采用该模型进行蓄水坑灌果园土壤贮水量预测。
This paper aims to realize the accurate prediction of soil moisture content under water storage pit irrigation. Based on the least squares support vector machine, BP neural network and the BP neural network optimized by the incremental back propagation algorithm, the LSSVM-WSP model, BP-WSP model and IBP-BP-WSP model were established. The initial moisture content, forecast period length, irri-gation amount in period, rainfall in period, and evapotranspiration were taken as input, the soil water storage in the root zone, which was treated with water storage pit irrigation, were taken as output. The LSSVM-WSP model, BP-WSP model and IB P-BP-WSP model were val-idated by field test data. The results show that the mean absolute percentage error of LSSVM-WSP model, BP-WSP model and IB P -BP- WSP model is 6.53%,3.64% and 5.98% respectively. The IB P-BP-WSP has the best precision, and it can be used to predict the soil mois-ture storage under water storage pit irrigation.
出处
《人民黄河》
CAS
北大核心
2017年第10期145-148,共4页
Yellow River
基金
国家自然科学基金资助项目(51579168,51249002)
山西省自然科学基金资助项目(201601D011053)
山西省高等学校创新人才支持计划项目
山西省科技攻关项目
关键词
土壤贮水量
最小二乘向量机
增量逆传播
BP神经网络模型
T检验
soil water storage
least squares vector machine
incremental back propagation algorithm
BP neural network
t-test