基于BP神经网络的棉花颜色级预测

为了方便快捷检验棉花颜色级,采用BP神经网络算法构建了基于棉纤维色度参数的棉花颜色级预测模型。模型以棉纤维的色度参数明度值和黄度值为输入变量,以棉花颜色级为输出变量,采用三层网络,网络的拓扑结构为2-12-1。通过对280份棉花试样的颜色级进行测试和BP神经网络模型预测,结果表明:BP神经网络预测值与实测值的平均相对误差为3.203%,所建立的回归方程拟合优度达到99%。认为:所建立BP神经网络模型具有较好的预测效果。     

用BP神经网络预测精纺毛织物后整理质量

通过确定合理的输入层单元数、隐层节点数,以3层BP神经网络建立了原料、纱线和后整理工艺参数与织物质量间的关系,对实际加工工艺进行预测,获得较满意的效果,证明该方法的实用性;同时建立了织物质量与工艺参数间的关系,并对重要工艺参数进行反演预测,为控制生产提供了指导.     

基于粒子群优化算法的坯布等级预测研究

本文简单介绍了织机工艺参数的特点及其对坯布质量的影响。建立了输出为坯布等级、输入为织机工艺参数的神经网络模型。最后,选用粒子群优化算法训练该神经网络模型,合理的确定了该神经网络的权值和阈值。     

模糊神经网络应用于食用果酱的可行性研究

研究了pH值、萃取温度、萃取时间和样品中氯化钠浓度等不同参数萃取条件对果酱中防腐剂萃取效果的影响,并建立了模糊神经网络模型对果酱中的防腐剂种类进行预测。结果表明:在pH值为3.3,萃取温度为室温,样品中氯化钠浓度为34%,萃取时间达到90min时,果酱中的防腐剂萃取效果最好。通过建立模糊神经网络模型,模型结构参数为输入层4,输出层5,最小训练速率0.1,动态参数0.6。对果酱中防腐剂的类型进行预测,试验结果表明模糊神经网络模型可以用于预测果酱中所含防腐剂的种类。     

苹果渣多酚提取工艺的神经网络建模与遗传算法优化

本实验建立了苹果渣多酚提取工艺的人工神经网络模型,研究了提取工艺神经网络模型的遗传算法优化技术。结果表明,结构为4-9-1的神经网络能较为精确的拟合输入的样本数据,其对测试样本的输出与实验结果的相关系数为0.985;遗传算法优化出的最佳提取工艺参数为乙醇浓度为62%、乙醇溶液的体积与苹果渣的质量之比为14:1,温度为69.7℃,提取时间为5.9h,该工艺参数下的提取率明显大于单因素试验和二次组合试验的结果,比最好的大16.9%。用神经网络模型描述提取工艺参数与提取率之间的关系,用遗传算法优化工艺参数,能设计出最佳的提取工艺参数。     

以客观指标为输入参数的织物热湿舒适性预测

探讨以客观指标为输入参数的织物热湿舒适性预测效果.测试了36种针织试样的6个静态客观热湿舒适性评价指标以及4个主观热湿舒适性评价指标,利用BP神经网络的方法建立了以静态客观评价指标为输入参数、以主观评价指标为输出参数的织物热湿舒适性预测模型.应用该模型对其中8种针织试样进行检验和分析,预测结果表明:该模型能有效地预测织物的主观热湿舒适性,其预测绝对误差范围在0.04 ～1.25间.     

毛精纺前纺工艺参数重要性的BP网络定量评价法

在BP神经网络建模技术的基础上,提出利用神经网络输入层与输出层之间的网络权值及其分布来求各输入参数重要程度的方法。将采集到的毛精纺企业前纺工艺参数运用BP神经网络分别建立了粗纱CV值和粗纱单重的预测模型。结果表明:所建模型的平均相对误差都低于3%;采用样本数据验证,其预报值与实测值间的相关系数都高于0.95。对所建模型的网络权重进行提取,分别计算出13个输入参数对粗纱CV值和粗纱单重的重要性,挖掘出显著而有效的参数。经对比认为,BP网络法比多元回归显著性分析(MRSA)更为精准,可用于对实际生产加工的预报和控制。     

基于BP神经网络的口罩防护舒适性研究

冬季雾霾天气频现,然而可供选择的口罩很少。探讨了以客观指标作为输入参数的口罩防护舒适性的预测结果。通过测试16种口罩试样的5个客观评价指标以及4个主观防护舒适性评价指标,建立预测口罩防护舒适性的BP神经网络模型。以客观测试指标为模型的输入参数,以主观评价指标为模型的输出参数,并在16组试样中随机选取4组来判断模型的准确性。结论显示:模型对防霾口罩的防护舒适性预测具有一定的可靠性,其预测绝对误差范围在0.5~0.8。     

服装智能制板中的结构特征参数模型设计

为了简化服装制板中结构特征参数的获取过程,建立款式、结构特征参数的映射关系。基于服装款式、结构的特征,以西装领为研究对象,通过二维款式图、结构图确定其特征参数,并以PyTorch深度学习框架为基础,利用BP神经网络算法建立预测模型,拟合款式特征参数与结构特征参数的非线性关系,最终均方误差达到0.000 019,完成了服装结构特征参数的智能化确定过程。利用55组测试数据,对该模型进行测试,其测试结果表明:55组测试数据的相对误差均在-0.04%～0.04%之间,预测模型精确度较高。利用该预测模型,通过给定的款式特征参数能得到与其映射的结构特征参数,将所获参数输入到自动绘图系统中即得到相应的样板。该模型的设计有利于提高服装制板速度,减少制板师工作量。     

基于非线性偏最小二乘的操作臂逆解算法

操作臂逆运动学问题是机器人控制中的一项重要内容。目前使用较多的神经网络法大多为多输入多输出或者多输入单输出方式,需要大量运算。非线性偏最小二乘法(NLPLS)建立的模型分为内部和外部模型,样本数据经外部模型处理后才用于训练若干个单输入单输出的神经网络。对PUMA560操作臂的仿真试验表明,在相同隐层神经元数的情况下,该算法比普通神经网络法具有更好的预测精度。这也表明,NLPLS只需较少的隐层神经元数就可以达到普通方法的精度,从而减少运算量。     

近红外光谱协同BP神经网络的泰国茉莉香米掺伪含量快速鉴定

利用近红外光谱协同BP神经网络算法,对泰国茉莉香米及其掺伪样品的近红外光谱进行多元散射校正预处理,挑选出48个特征波长;以特征波长的吸光度为BP神经网络输入层神经元,以样品中泰国茉莉香米的含量为输出层神经元,获得BP神经网络算法的最优结构模型,即为单层隐含层、隐含层神经元数7、隐含层传递函数logsig、输出层传递函数...     

基于BP神经网络的运动文胸肩带属性与乳房振幅的函数关系

为确定运动文胸肩带的3种属性在人体跑步时对胸部振幅的影响,选取8名被测人员,在其左胸上标记6个测量点,更换不同的肩带进行人体运动测试,记录这些测量点动态的三维坐标,进而得到乳房运动的振幅;利用BP神经网络模型,通过更换不同的网络模型参数,确定运动文胸肩带的3种属性与乳房振幅之间的权值关系。结果表明,选取BP神经网络的传输函数为tansig函数,隐含层神经元个数为21个,训练函数为traingdm作为网络参数时,网络拟合出的乳房振幅值达到了真实值的99.44%;在该网络参数下,分别求得网络输入层到隐含层和隐含层到输出层的权值和阈值,最终得到肩带的3种属性与胸部振幅的正向推理关系式。     

Modeling of acrylamide formation and browning ratio in potato chips by artificial neural network

The artificial neural network (ANN) modeling approach was used to predict acrylamide formation and browning ratio (%) in potato chips as influenced by time x temperature covariants. A series of feed-forward type network models with back-propagation training algorithm were developed. Among various network configurations, 4-5-3-2 configuration was found as the best performing network topology. Four neurons in the input layer were reflecting the asparagine concentration, glucose concentration, frying temperature, and frying time. The output layer had two neurons representing acrylamide concentration and browning ratio of potato chips. The ANN modeling approach was shown to successfully predict acrylamide concentration (R = 0.992) and browning ratio (R = 0.997) of potato chips during frying at different temperatures in time-dependent manner for potatoes having different concentrations of asparagine and glucose. It was concluded that ANN modeling is a useful predictive tool which considers only the input and output variables rather than the complex chemistry.     

Predicting the drying kinetics of salted codfish (Gadus Morhua): semi‐empirical,diffusive and neural network models

This work aims to compare the accuracy of several drying modelling techniques namely semi‐empirical, diffusive and artificial neural network (ANN) models as applied to salted codfish (Gadus Morhua). To this end, sets of experimental data were collected to adjust parameters for the models. Modelling of codfish drying was performed by resorting to Page and Thompson semi‐empirical models and to a Fick diffusion law. The ANN employed a neural network multilayer ‘feed‐forward’, consisting of one input layer, with four neurons, one hidden layer, formed by five neurons and one output layer with a convergence criterion for training purposes. The simulations showed good results for the ANN (correlation coefficient between 0.987 and 0.999) and semi‐empirical models (correlation coefficient ranging from 0.992 to 0.997 for Page’s model, and from 0.993 to 0.996 for Thompson’s model), while improvements were required to obtain better predictions by the diffusion model (correlation coefficients ranged from 0.864 to 0.959).     

Artificial neural network modeling of physicochemical changes of shrimp during boiling

Frozen boiled shrimp and dried shrimp are among the high-value fishery products of Thailand. During the production of these products boiling is one of the most important steps that affects significantly the product physicochemical properties, especially the quantity and quality of proteins, which in turn affect other apparent properties perceived by consumers. The protein changes are, however, difficult to evaluate comparing to other typical physical properties of shrimp. The objective of this study was therefore to develop an artificial neural network (ANN) model to predict the protein changes of shrimp in terms of protein loss and protein denaturation as a function of the boiling conditions, namely, concentration of salt solution and boiling time, as well as a rather easily determined change of shrimp, that is, cooking loss. Other apparent physical changes of shrimp viz. shrinkage and hardness were also predicted. The optimized ANN structure for the prediction of protein changes consists of two hidden layers and twelve neurons per layer, while the optimized ANN structure for the prediction of physical changes consisted of two hidden layers and eight neurons per layer. The predicted protein and physical changes agreed very well with the experimental data with R² > 0.994 and 0.972, respectively.     

基于改进ResNet模型的食品新鲜度识别方法

目的:解决现有食品新鲜度识别方法存在的检测效率低和精度差等问题。方法:基于食品生产线图像采集系统,提出一种改进的残差神经网络模型用于生产线食品新鲜度识别。引入改进的LRELU激活函数提高模型的识别性能,引入批量归一化层提高模型的训练效率,引入Dropout层丢弃一定比例的神经元降低过拟合的影响。结果:与常规食品新鲜度识别方法相比,试验方法能够较为准确、高效地实现食品新鲜度识别,总体新鲜度识别准确率>97%,平均识别时间为9.8 ms,满足食品生产线对新鲜度识别的需要。结论:基于深度学习的检测方法是一种无损、高效、高精度的食品图像新鲜度识别方法。     

第三层交换技术及其在校园网中的应用

第三层交换技术是将网络层的功能融入交换机,将第二层交换机与第三层路由器两者的优势集成在一起,在各个层次提高网络速度性能的一种网络技术。用三层交换机可取代传统路由器充当局域网路由器的作用,采用三层交换技术还可有效地防止ＩＰ盗用,提高网络的安全性,并使网络管理员对网络管理更加灵活方便。因此,三层交换技术可成为中小型校园网建设的有效解决方案。     

Artificial neural network modeling and optimization of ultrahigh pressure extraction of green tea polyphenols

In this study, the ultrahigh pressure extraction of green tea polyphenols was modeled and optimized by a three-layer artificial neural network. A feed-forward neural network trained with an error back-propagation algorithm was used to evaluate the effects of pressure, liquid/solid ratio and ethanol concentration on the total phenolic content of green tea extracts. The neural network coupled with genetic algorithms was also used to optimize the conditions needed to obtain the highest yield of tea polyphenols. The obtained optimal architecture of artificial neural network model involved a feed-forward neural network with three input neurons, one hidden layer with eight neurons and one output layer including single neuron. The trained network gave the minimum value in the MSE of 0.03 and the maximum value in the R² of 0.9571, which implied a good agreement between the predicted value and the actual value, and confirmed a good generalization of the network. Based on the combination of neural network and genetic algorithms, the optimum extraction conditions for the highest yield of green tea polyphenols were determined as follows: 498.8 MPa for pressure, 20.8 mL/g for liquid/solid ratio and 53.6% for ethanol concentration. The total phenolic content of the actual measurement under the optimum predicated extraction conditions was 582.4 ± 0.63 mg/g DW, which was well matched with the predicted value (597.2 mg/g DW). This suggests that the artificial neural network model described in this work is an efficient quantitative tool to predict the extraction efficiency of green tea polyphenols.     

基于量子卷积神经网络算法的微小零件识别

设计了量子卷积神经网络表示层、隐藏层神经元和输出层神经元模型；采用修正线性激活函数ReLu作为激活函数，并通过训练误差函数优化量子旋转角度和神经连接权值。8种微小零件的仿真试验表明，量子卷积神经网络算法的识别准确率较高，耗时少且识别效果较好。     

Modeling Thermal Conductivity of Iranian Flat Bread Using Artificial Neural Networks

The objective of this work was to develop Artificial Neural Network (ANN) based thermal conductivity (K) prediction model for Iranian flat breads. Experimental data needed for ANN models were obtained from a pilot-scale set-up. Breads were made from three different cultivars of wheat and were baked in an eclectic oven at three different baking temperatures (232°C, 249°C and 260°C). A data set of 205 conditions was used for developing ANN and empirical models. To model K using ANN, 16 different MLP (multilayer perceptron) configurations ranging from one to two hidden layers of neurons were investigated and their prediction performances were evaluated. The (4-3-5-1)-MLP network, that is a network having two hidden layers, with three neurons in its first hidden layer and five neurons in its second hidden layer, had the best results in predicting the thermal conductivity of flat bread. For this network, R², MRE, MAE and SE were 0.988, 0.6323, 1.66×10^{? 3}, and 8.56×10^?4, respectively. Overall, ANN models (with R² ≥ 0.95) performed superior than the empirical model (with R² = 0. 870).