论文部分内容阅读
作物参考蒸腾量(ET0)是作物生长过程中一个非常重要的数据,ET0反应的是大气蒸发能力与作物需水信息的关系。采用自适应神经模糊推理系统(ANFIS)研究易于获得的日最低气温、日最高气温、日平均气温、日平均相对湿度、实际日照时长及风速六项气象数据对作物参考蒸腾量(ET0)的相关程度,通过比较均方根误差找到相关程度最大的组合且结构简单的ANFIS模型来预测ET0值,并通过比较均方根误差来验证所建立的ANFIS模型预测的准确性。结果表明,综合考虑模型的预测精度及结构的复杂程度,日最高气温、日平均相对湿度和风速的三输入组合为最佳的,其平均绝对误差小且ANFIS结构简单。利用该输入组合训练的ANFIS模型预测ET0,其训练的均方根误差相比于用神经网络训练的预测模型小,通过比较可知ANFIS比BP神经网络训练的模型精度提高。
Crop reference evapotranspiration (ET0) is a very important data in crop growth. ET0 reflects the relationship between atmospheric evapotranspiration and crop water requirement information. An adaptive neuro-fuzzy inference system (ANFIS) was used to study the correlations of available daily minimum temperature, daily maximum temperature, daily average temperature, daily average relative humidity, actual sunshine duration and wind speed on crop reference transpiration (ET0) The ANFIS model with the most relevant correlation and the simple structure was found by comparing the RMSE to predict the ET0 value. The accuracy of the ANFIS model prediction was verified by comparing with the root mean square error. The results show that the three-input combination of daily maximum air temperature, daily average relative humidity and wind speed is the best, and the average absolute error is small and the structure of ANFIS is simple, considering the prediction accuracy and structural complexity of the model. Compared with the prediction model trained by neural network, the root mean square error of training forecast ET0 by using the input training ANFIS model. By comparing, it can be seen that the accuracy of ANFIS training model is improved than that of BP neural network training.