ANN models

The ANN model had four neurones in the input layer one for each operating variable and one for the bias. The output was selected to be cumulative mass distribution thirteen neurones were used to represent it. A sigmoid functional... 

The computational paradigm in an ANN is based on an idealized model of a biological unit called a neuron. The unique characteristics of this ANN model are the inputs of signal from stimulus in a training environment. It is important to note that each neuron works independently of the other neurons. The specific characteristics of ANN models that attract industrial application are ... 

Figure 9 Comparison of viscosity predictions between ANN model and Batschinski s equation.

Figure 25 ANN model (5-8-6) training and testing results for van der Waals volume, molar volume, heat capacity, solubility parameter, and glass transition temperature of 45 different polymers.

Figure 30 ANN model training and testing results for crosslink density changes due to aging.

Bourquin J, Schmidli H, van Hoogevest P, Leuenberger H. Basic concepts of artificial neural networks (ANN) modelling in the application to pharmaceutical development. Pharm Dev Technol 1997 2 95-109. 

For PyMS to be used for (1) routine identification of microorganisms and (2) in combination with ANNs for quantitative microbiological applications, new spectra must be comparable with those previously collected and held in a data base.127 Recent work within our laboratory has demonstrated that this problem may be overcome by the use of ANNs to correct for instrumental drift. By calibrating with standards common to both data sets, ANN models created using previously collected data gave accurate estimates of determi-nand concentrations, or bacterial identities, from newly acquired spectra.127 In this approach calibration samples were included in each of the two runs, and ANNs were set up in which the inputs were the 150 new calibration masses while the outputs were the 150 old calibration masses. These associative nets could then by used to transform data acquired on that one day to data acquired at an earlier data. For the first time PyMS was used to acquire spectra that were comparable with those previously collected and held in a database. In a further study this neural network transformation procedure was extended to allow comparison between spectra, previously collected on one machine, with spectra later collected on a different machine 129 thus calibration transfer by ANNs was affected. Wilkes and colleagues130 have also used this strategy to compensate for differences in culture conditions to construct robust microbial mass spectral databases. 

Both cases can be dealt with both by supervised and unsupervised variants of networks. The architecture and the training of supervised networks for spectra interpretation is similar to that used for calibration. The input vector consists in a set of spectral features yt(Zj) (e.g., intensities at selected wavelengths zi). The output vector contains information on the presence and absence of certain structure elements and groups fixed by learning rules (Fig. 8.24). Various types of ANN models may be used for spectra interpretation, viz mainly such as Adaptive Bidirectional Associative Memory (BAM) and Backpropagation Networks (BPN). The correlation... 

Development of ANN model for estimating VLE is less cumbersome than methods based on EOS. It does not require parameters such as the critical properties of the components or the... 

Degim T, Hadgraft J, Ilbasmis S, Ozkan Y (2003) Prediction of skin penetration using artificial neural network (ANN) modeling. J Pharm Sci 92 656-664. 

Yamamura S (2003) Clinical application of artificial neural network (ANN) modeling to predict pharmacokinetic parameters of severely ill patients. Adv Drug Deliv Rev 5 1233-1251. 

Agantonovic-Kastrin S, Beresford R, Yusof AP (2001) ANN modeling of the penetration across a polydimethylsiloxane membrane from theoretically derived molecular descriptors. J Pharm Biomed Anal 26 241-254. 

Table 12.7 The model fit (RMSEE) values of 7 different neural net (ANN) models for predicting cis-butadiene content in styrene-butadiene copolymers by NIR spectroscopy using 1 to 6 nodes in the hidden layer...

Unfortunately, the ANN method is probably the most susceptible to overfitting of the methods discussed thus far. For similar N and M, ANNs reqnire many more parameters to be estimated in order to define the model. In addition, cross validation can be very time-consuming, as models with varying complexity (nnmber of hidden nodes) mnst be trained individually before testing. Also, the execntion of an ANN model is considerably more elaborate than a simple dot product, as it is for MLR, CLS, PCR and PLS (Eqnations 12.34, 12.37, 12.43 and 12.46). Finally, there is very little, or no, interpretive value in the parameters of an ANN model, which eliminates one nseful means for improving the confidence of a predictive model. 

For the styrene-butadiene copolymer application, fit results of ANN models using one to six hidden nodes are shown in Table 12.7. Based on these results, it appears that only three, or perhaps four, hidden nodes are required in the model, and the addition of more hidden nodes does not greatly improve the fit of the... 

M. Blanco and A. Peguero, An expeditious method for determining particle size distribution by near infrared spectroscopy comparison of PLS2 and ANN models, Talanta, 77(2), 647-651 (2008). 

Yu X, Yi B, Liu F et al. (2008) Prediction of the dielectric dissipation factor tan delta of polymers with an ANN model based on DFT calculation. React Fund Polym 68 1557-1562... 

Electronic descriptors were calculated for the ab initio optimized (RHG/STO-3G) structures. In addition, logP as a measure of hydrophobicity and different topological indices were also calculated as additional descriptors. A nonlinear model was constructed using ANN with back propagation. Genetic algorithm (GA) was used as a feature selection method. The best ANN model was utilized to predict the log BB of 23 external molecules. The RMSE of the test set was only... 

In order to develop an ANN model for a refinery unit that is able to predict product yields and properties, one must first decide on the important inputs and outputs of the process. The choice of these inputs and outputs is the most important factor in successfully preparing an ANN model. These inputs and outputs must be chosen by carefully examining plant data. A good expertise in relation to the process is necessary. 

Another important factor in developing an ANN model is deciding the architecture of the network. This decision is often made on a trial and error basis. First, a network with one hidden layer only and with a fixed number of neurons is chosen. 

In order to develop an ANN model for the FCC process, we use here the same data set as in the previous section (Section 2.4). This data set was divided into two sets, one set for training and one set for testing the neural network. The prepared network model is able to predict the yields of the various FCC products and also the CCR number. During training of the neural network, first, only one hidden layer with five neurons was used. This network did not perform well against a pre-specified tolerance of 10-3. 

As can be seen from the tables, the ANN model consistently gives better predictions. Themainreasonis that the simulator required a lot of input information which had to be estimated while the neural network model required only four feed properties. 

Table 2.12 Comparison of ANN models and simulation results for coke and gasoline yields.

Table 2.13 Comparison of ANN models and simu lation results foi coke and gasoline. ...

---