In this study a neural networks model designed in Matlab 7.0 was used to predict the conditions of the tray drying, the osmotic dehydration and cooking-dehydration of vegetal pear (Sechium edule) slices. In this way this study verifies the influence of the variables, temperature (53-73 °C), feed material (3-9 kg) and diameter of the slices (4-6 cm), through a 2~3 x 3 design, in order to determine the moisture, water activity, color difference, production (kg h~(-1)) and cost ($ kg~(-1)) on the drying process. The optimal configuration of the neural network model was obtained by varying the main parameters as transfer function, learning rate, interactions number, hidden neuron, normalization interval, number of data for training and for test. The correlation coefficients between the training data set and the predictions of the model exceeded 0.97. In general, the results indicated that it is feasible to use the neural network model for prediction and optimization of the three different drying process of vegetal pear.
展开▼
机译：在本研究中，使用在Matlab 7.0中设计的神经网络模型来预测纸盘干燥，渗透脱水和植物梨（Sechium Edule）切片的渗透脱水的条件。通过这种方式，本研究验证了变量，温度（53-73°C），饲料材料（3-9千克）和切片直径（4-6厘米）的影响，通过2〜3 x 3设计，为了确定干燥过程中的水分，水活动，色差，生产（kg h〜（-1））和成本（kg〜（-1））。通过改变主要参数作为传递函数，学习速率，交互号码，隐藏的神经元，归一化间隔，培训数据数量来获得神经网络模型的最佳配置。训练数据集之间的相关系数和模型的预测超过0.97。通常，结果表明，使用神经网络模型进行预测和优化植物梨的三种不同干燥过程是可行的。
展开▼
