基于MATLAB的温室番茄经济效益最优目标研究与预测

Prediction of the impact of environmental conditions on greenhouse tomato input-output ratio based on MATLAB( matrix laboratory)

以温室番茄产出与投入比值作为目标进行正交试验,研究温室内番茄生长的环境参数对产投比的影响规律,建立BP神经网络模型。运用MATLAB对试验数据进行训练和模拟,采用10-折交叉验证,检验预测的可靠性,准确率为93.7%。表明利用BP神经网络得出的预测值与实测值接近,具有较好的预测性,可用于产投比预测,能够为温室环境调控提供理论依据。

The orthogonal experiments on greenhouse tomato were conducted to analyze the impact of environmental parameters on input/ output. BP neural network model was established, and the data was trained and simulated by MATLAB（matrix laboratory）. Through 10-fold cross validation, model achieved the predictive accuracy of 93. 7%, in good agreement with the experimental values. BP neural network model is feasible in precision of input/ output of tomato.