Prediction of the Profit Function for Industrial 2‐Keto‐L‐Gulonic Acid Cultivation

The profit function is the generic criterion to describe the cost effect of a batch process. To focus on the prediction of the profit function for 2‐keto‐L‐gulonic acid (2‐KGA) cultivation, which is potentially applicable for process monitoring and optimal scheduling, rolling learning‐prediction (RLP) based on a support vector machine (SVM) is applied. The RLP implies that the SVM training database is rolling updated as the batch of current interest proceeds, and the SVM learning is then repeated for the prediction. The database is further updated after termination of a batch. The updating procedures are investigated in detail. Pseudo‐online prediction is carried out using the data from industrial‐scale 2‐KGA cultivation under actual and hypothetical inoculation sequences. The results indicate that the average relative prediction error is less than 5 % in the later phase of fermentation in all inoculation sequences.     

Artificial neural networks for colour prediction in leather dyeing on the basis of a tristimulus system

Computer‐assisted colour prediction and quality control have become increasingly important to the dyeing process in many consumer goods manufacturing industries, including textile and leather. The most challenging aspect concerns dye recipe prediction for the production of the required shade on a given substrate. Computer recipe prediction based on the conventional and widely used Kubelka–Munk model often fails under a variety of conditions. In the present investigation, an attempt has been made to develop an artificial neural network model to predict colour in terms of tristimulus values (X, Y, Z) given the concentration of dyes. An artificial neural network model was trained with 300 pairs of known input vectors, i.e. dye concentrations, and output vectors, i.e. colour parameters, using a backpropagation algorithm. The artificial neural network topology consists of three neurons in the input layer to represent the concentration of dyes, three neurons in the output layer to represent the tristimulus values X, Y, and Z, and five neurons in the hidden layer with a log‐sigmoid transfer function. The artificial neural network results showed a good level of colour prediction during the training and testing phase. The results also indicate that the artificial neural network has the potential to give better predictive performance than the conventional Kubelka–Munk model.     

基于向量投影的代谢支持向量机乙烯精馏产品质量软测量建模

针对支持向量机(SVM)增量学习过程中易出现计算速度慢、稳定性差的缺陷,提出了一种基于向量投影的代谢支持向量机建模方法.该方法首先运用向量投影算法对训练样本进行预选取来减少样本数量,提高SVM建模速度.然后将新增样本"代谢"原则引入SVM增量学习过程中,以解决因新增样本不断加入而导致训练样本数量"爆炸"的问题.最后将该方法用于乙烯精馏产品质量软测量建模,实验结果表明,与传统SVM和最小二乘支持向量机(LSSVM)相比,向量投影的代谢SVM具有更好的预测结果.     

Development of calibration models for quality control in the production of ethylene/propylene copolymers by FTIR spectroscopy,multivariate statistical tools,and artificial neural networks

Principal component regression (PCR), partial least squares (PLS), StepWise ordinary least squares regression (OLS), and back‐propagation artificial neural network (BP‐ANN) are applied here for the determination of the propylene concentration of a set of 83 production samples of ethylene–propylene copolymers from their infrared spectra. The set of available samples was split into (a) a training set, for models calculation; (b) a test set, for selecting the correct number of latent variables in PCR and PLS and the end point of the training phase of BP‐ANN; (c) a production set, for evaluating the predictive ability of the models. The predictive ability of the models is thus evaluated by genuine predictions. The model obtained by StepWise OLS turned out to be the best one, both in fitting and prediction. The study of the breakdown number of samples to be included in the training set showed that at least 52 experiments are necessary to build a reliable and predictive calibration model. It can be concluded that FTIR spectroscopy and OLS can be properly employed for monitoring the synthesis or the final product of ethylene–propylene copolymers, by predicting the concentration of propylene directly along the process line. © 2008 Wiley Periodicals, Inc. J Appl Polym Sci, 2008     

径向基函数神经网络在砂体厚度预测中的应用

目前,人工智能神经网络在地震储层参数的预测方面具有广泛的应用,最常用的为BP神经网络,但是效果并不是十分理想。径向基函数神经网络(RBFN)是一种前馈神经网络,其在函数逼近、模式识别方面都优于BP网络,已经在岩性识别、孔渗预测方面取得了较好的应用效果。本文首次将此方法运用于预测砂体厚度,利用地震属性信息和神经网络的学习,基于实际数据计算,最后计算出相应的砂体厚度值,并与实测值进行误差分析。实例分析表明,利用径向基函数神经网络进行砂体厚度预测具有一定的可行性和实用价值。     

Machine learning-based predictive control of nonlinear processes. Part I: Theory

This article focuses on the design of model predictive control (MPC) systems for nonlinear processes that utilize an ensemble of recurrent neural network (RNN) models to predict nonlinear dynamics. Specifically, RNN models are initially developed based on a data set generated from extensive open-loop simulations within a desired process operation region to capture process dynamics with a sufficiently small modeling error between the RNN model and the actual nonlinear process model. Subsequently, Lyapunov-based MPC (LMPC) that utilizes RNN models as the prediction model is developed to achieve closed-loop state boundedness and convergence to the origin. Additionally, machine learning ensemble regression modeling tools are employed in the formulation of LMPC to improve prediction accuracy of RNN models and overall closed-loop performance while parallel computing is utilized to reduce computation time. Computational implementation of the method and application to a chemical reactor example is discussed in the second article of this series.     

Colour‐appearance modeling using feedforward networks with Bayesian regularization method. Part II: Reverse model

In Part I of this article, the development of a multilayer perceptrons feedforward artificial neural network model to predict colour appearance from colorimetric values was reported. Bayesian regularization was employed for the training of the network. In this part of the article, the reverse model, that is, the perdition of colorimetric values from the colour appearance attributes is reported using the same neural network design methodology developed in Part I. This study should contribute to the building of an artificial neural network–based colour appearance prediction, both forward and reverse, using the most comprehensive LUTCHI colour appearance data sets for training and testing. Good prediction accuracy and generalization ability were obtained using the neural networks built in the study. Because the neural network approach is of a black‐box type, colour appearance prediction using this method should be easier to apply in practice. © 2002 Wiley Periodicals, Inc. Col Res Appl, 27, 116–121, 2002; DOI 10.1002/col.10030     

基于最小二乘支持向量机的天然气负荷预测

对城市天然气负荷预测的研究,对于保证天然气管网用气量、优化管网的调度和设备维修具有极其重要的意义．在国内,对于城市天然气负荷预测的研究才刚刚起步,目前还没有较系统的理论．同技术与理论较为成熟的电力负荷预测研究相比较,两者既有许多相同点,又有不同之处．相同之处在     

TransPhos: A Deep-Learning Model for General Phosphorylation Site Prediction Based on Transformer-Encoder Architecture

Protein phosphorylation is one of the most critical post-translational modifications of proteins in eukaryotes, which is essential for a variety of biological processes. Plenty of attempts have been made to improve the performance of computational predictors for phosphorylation site prediction. However, most of them are based on extra domain knowledge or feature selection. In this article, we present a novel deep learning-based predictor, named TransPhos, which is constructed using a transformer encoder and densely connected convolutional neural network blocks, for predicting phosphorylation sites. Data experiments are conducted on the datasets of PPA (version 3.0) and Phospho. ELM. The experimental results show that our TransPhos performs better than several deep learning models, including Convolutional Neural Networks (CNN), Long-term and short-term memory networks (LSTM), Recurrent neural networks (RNN) and Fully connected neural networks (FCNN), and some state-of-the-art deep learning-based prediction tools, including GPS2.1, NetPhos, PPRED, Musite, PhosphoSVM, SKIPHOS, and DeepPhos. Our model achieves a good performance on the training datasets of Serine (S), Threonine (T), and Tyrosine (Y), with AUC values of 0.8579, 0.8335, and 0.6953 using 10-fold cross-validation tests, respectively, and demonstrates that the presented TransPhos tool considerably outperforms competing predictors in general protein phosphorylation site prediction.     

A novel approach to using neural networks to predict the colour of fibre blends

This work is concerned with the colour prediction of viscose fibre blends, comparing two conventional prediction models (the Stearns–Noechel model and the Friele model) and two neural network models. A total of 333 blended samples were prepared from eight primary colours, including two‐, three‐, and four‐colour mixtures. The performance of the prediction models was evaluated using 60 of the 333 blended samples. The other 273 samples were used to train the neural networks. It was found that the performance of both neural networks exceeded the performance of both conventional prediction models. When the neural networks were trained using the 273 training samples, the average CIELAB colour differences (between measured and predicted colour of blends) for the 60 samples in the test set were close to 1.0 for the neural network models. When the number of training samples was reduced to only 100, the performance of the neural networks degraded, but they still gave lower colour differences between measured and predicted colour than the conventional models. The first neural network was a conventional network similar to that which has been used by several other researchers; the second neural network was a novel application of a standard neural network where, rather than using a single network, a set of small neural networks was used, each of which predicted reflectance at a single wavelength. The single‐wavelength neural network was shown to be more robust than the conventional neural network when the number of training examples was small.     

基于支持向量机回归的天然气负荷预测研究

Natural gas load forecasting is a key process to the efficient operation of pipeline network. An accurate forecast is required to guarantee a balanced network operation and ensure safe gas supply at a minimum cost. Machine learning techniques have been increasingly applied to load forecasting. A novel regression technique based on the statistical learning theory, support vector machines (SVM), is investigated in this paper for natural gas shortterm load forecasting. SVM is based on the principle of structure risk minimization as opposed to the principle of empirical risk minimization in conventional regression techniques. Using a data set with 2 years load values we developed prediction model using SVM to obtain 31 days load predictions. The results on city natural gas short-term load forecasting show that SVM provides better prediction accuracy than neural network. The software package natural gas pipeline networks simulation and load forecasting (NGPNSLF) based on support vector regression prediction has been developed, which has also been applied in practice.     

Prediction of Timber Kiln Drying Rates by Neural Networks

The purpose of this exploratory work was to apply artificial neural network (ANN) modeling to the prediction of timber kiln drying rates based on species and basic density information for the hem-fir mix that grows along the local coastal areas. The ANN models with three inputs (initial moisture content, basic density, and drying time) were developed to predict one output, namely, average final moisture content. The back-propagation algorithm, the most common neural network learning method, was implemented for testing, training, and validation. Optimal configuration of the network model was obtained by varying its main parameters, such as transfer function, learning rule, number of neurons and layers, and learning runs. Accurate prediction of the experimental drying rate data by the ANN model was achieved with a mean absolute relative error less than 2%, thus supporting the powerful predictive capacity of this modeling method.     

The Development of Intelligent Models for Liquid–Liquid Equilibria (LLE) Phase Behavior of Thiophene/Alkane/Ionic Liquid Ternary System

ABSTRACT

This paper develops three models based on artificial neural network (ANN), support vector machine (SVM) and least square support vector machine (LSSVM) algorithm for phase behavior of thiophene/alkane/ionic liquid ternary system. The shuffled complex evolution (SCE) was employed to acquire the optimal magnitudes of hyper parameters (σ² and γ) which are embedded parts of SVM and LSSVM models, and the trial and error was employed to obtain the optimal numbers of neuron and layers for ANN intelligent model. Gathering and using 618 LLE data, the comparison between the optimized version of applied intelligent models in giving the LLE was also made. The findings are indicative of a prefect agreement between the estimation from intelligent models and the experimental data. The finding also reveals that the performance of SVM in prediction of solubility is somewhat better than other intelligent models (i.e., ANN and SVM) as coefficient determination (R2) and root mean squared error (RMSE) are respectively 0.9961 and 0.0447 for test sets of data. This is likely due to the existence of structural risk minimization principle of SVM which is embodied in SVM algorithm and effectively minimizes upper bound of the generalization error, rather than minimizing the training error.     

Melt index prediction by RBF neural network with an ICO‐VSA hybrid optimization algorithm

Melt index (MI) is considered as one of the most significant parameter to determine the quality and the grade of the practical polypropylene polymerization products. A novel ICO‐VSA‐RNN (RBF neural network with ICO‐VSA algorithm) MI prediction model is proposed based on radial basis function (RBF) neural network and improved chaos optimization (ICO), and variable‐scale analysis (VSA), where the ICO is first added and then combined with the VSA to overcome the defects of ICO and VSA, then the parameters of the RBF neural network are optimized with them. At last, the RBF neural network model for MI prediction model is developed. Further researches on the optimal RBF neural network model of MI prediction are carried out with the data from a real industrial plant, and the prediction results show that the performance of this prediction model is much better than the RBF neural network model without optimization. © 2012 Wiley Periodicals, Inc. J Appl Polym Sci, 2012     

基于BP神经网络紫外可见光度法测定苏丹红混合组分

应用人工神经网络,以BP算法对混合样品中苏丹红系列的三种组分(苏丹红Ⅰ、苏丹红Ⅲ、苏丹红Ⅳ)的浓度进行测定。在MATLAB 7.0中建立BP神经网络,优化网络条件,对训练样本进行训练,然后对检测样本进行检测。预测结果的误差范围在0.03%~9.20%之间。当样品浓度<0.1×10-4mol/L时,预测误差较大,均在5%以上;当浓度>0.1×10-4mol/L时,预测的相对误差较小,均在5%以下。该方法已用于模拟水样中微量苏丹红的检测。     

Enrichment of glutaminase production by Bacillus subtilis RSP‐GLU in submerged cultivation based on neural network—genetic algorithm approach

BACKGROUND: Owing to the importance of glutaminase in biotech product production, its production with isolated Bacillus subtilis RSP‐GLU (MTCC 9727) was investigated. Fermentation factors play an important role in product enhancement. Hence, glutaminase production was optimized using an artificial neural network (ANN) coupled genetic algorithm (GA). RESULTS: A ‘6–12–1’ topology ANN was constructed to identify the nonlinear relationship between fermentation factors and enzyme yield. ANN‐predicted values were optimized for glutaminase production using a GA. The overall mean absolute predictive error (MAPE) and the mean square errors (MSE) were observed to be 0.00125 and 1.77 and 0.002 and 3.06 for training and testing, respectively. The goodness of neural network prediction (coefficient of R²) was found to be 0.996. The maximum interactive impact on glutaminase production was noted with rpm versus medium volume. The use of ANN–GA hybrid methodology resulted in a significant improvement (47%) in glutaminase yield. CONCLUSION: Five different optimum fermentation conditions out of 500 revealed maximum enzyme production. Glutaminase enzyme production in this Bacillus subtilis RSP‐GLU is strongly influenced by aeration of the fermentation. A hybrid ANN‐GA effectively identifies the different fermentation conditions for optimum production of enzyme in a given large set of conditions. Copyright © 2009 Society of Chemical Industry     

Data-driven prediction model of indoor air quality in an underground space

Several data-driven prediction methods based on multiple linear regression (MLR), neural network (NN), and recurrent neural network (RNN) for the indoor air quality in a subway station are developed and compared. The RNN model can predict the air pollutant concentrations at a platform of a subway station by adding the previous temporal information of the pollutants on yesterday to the model. To optimize the prediction model, the variable importance in the projection (VIP) of the partial least squares (PLS) is used to select key input variables as a preprocessing step. The prediction models are applied to a real indoor air quality dataset from telemonitoring systems data (TMS), which exhibits some nonlinear dynamic behaviors show that the selected key variables have strong influence on the prediction performances of the models. It demonstrates that the RNN model has the ability to model the nonlinear and dynamic system, and the predicted result of the RNN model gives better modeling performance and higher interpretability than other data-driven prediction models.     

Research on Natural Gas Short-Term Load Forecasting Based on Support Vector Regression

Natural gas load forecasting is a key process to the efficient operation of pipeline network. An accurate forecast is required to guarantee a balanced network operation and ensure safe gas supply at a minimum cost. Machine learning techniques have been increasingly applied to load forecasting. A novel regression technique based on the statistical learning theory, support vector machines (SVM), is investigated in this paper for natural gas short-term load forecasting. SVM is based on the principle of structure risk minimization as opposed to the principle of empirical risk minimization in conventional regression techniques. Using a data set with 2 years load values we developed prediction model using SVM to obtain 31 days load predictions. The results on city natural gas short-term load forecasting show that SVM provides better prediction accuracy than neural network. The software package natural gas pipeline networks simulation and load forecasting (NGPNSLF) based on support vector regression prediction     

Prediction of Timber Kiln Drying Rates by Neural Networks

The purpose of this exploratory work was to apply artificial neural network (ANN) modeling to the prediction of timber kiln drying rates based on species and basic density information for the hem-fir mix that grows along the local coastal areas. The ANN models with three inputs (initial moisture content, basic density, and drying time) were developed to predict one output, namely, average final moisture content. The back-propagation algorithm, the most common neural network learning method, was implemented for testing, training, and validation. Optimal configuration of the network model was obtained by varying its main parameters, such as transfer function, learning rule, number of neurons and layers, and learning runs. Accurate prediction of the experimental drying rate data by the ANN model was achieved with a mean absolute relative error less than 2%, thus supporting the powerful predictive capacity of this modeling method.     

Derivation of optimal processing parameters of polypropylene foam thermoforming by an artificial neural network

The effects of processing parameters on the thermoforming of polymeric foam sheets are highly nonlinear and fully coupled. The complex interconnection of these dominant processing parameters makes the process design a difficult task. In this study, the optimal processing parameters of polypropylene foam thermoforming are obtained by the use of an artificial neural network. Data from tests carried out on a lab‐scale thermoforming machine were used to train an artificial neural network, which serves as an inverse model of the process. The inverse model has the desired product dimensions as inputs and the corresponding processing parameters as outputs. The structure, together with the training methods, of the artificial neural network is also investigated. The feasibility of the proposed method is demonstrated by experimental manufacturing of cups with optimal geometry derived from the finite element method. Except the dimension deviation at one location, which amounts to 17.14%, deviations of the other locations are all below 3.5%. POLYM. ENG. SCI., 45:375–384, 2005. © 2005 Society of Plastics Engineers