【摘要】 基于神经网络理论 ,对土壤施肥模型进行了研究 .通过对小麦在同一农田重复进行种植试验 ,且每次施不同的肥料用量 ,最后选取丰产的数据作为实验样本 ,通过归一化把实验数据进行必要的处理。采用BP算法训练网络 ,对小麦产量和施肥用量之间的映射关系进行了函数逼近 ,建立了土壤施肥神经网络模型 .通过实际数据验证 ,该模型的输出与实际肥料用量基本相符 .将所建神经网络模型运用于施肥量的预算 ,提高了预算精度并能取得较好的效果更多还原

【Abstract】 Based on the method of artificial neural network, the soil fertilizer model is researched. By repetition experiment of planting wheat in the same farmland and using the different dosage of fertilizer in every experiment, the data of fertility are finally selected and used as experiment sample. The data are dealt with by ANN. The artificial neural network is trained through adoption of BP algorithm.The map relationship between wheat yield and dosage of fertilizer is approached, and the soil fertilizer model is established. By practice verification, the actual output data of the model conforms to actual fertilizer dosage basically. Applying neural network model built to budget for fertilizer dosage can advance budget accuracy and can obtain good result.更多还原