Rapid optimization and minimal complexity in computational neural network multivariate calibration of chlorinated hydrocarbons using Raman spectroscopy

Egan, William J.; Angel, S. M.; Morgan, Stephen L.

doi:10.1002/1099-128x(200101)15:1<29::aid-cem600>3.0.co;2-a

Cited by 9 publications

(6 citation statements)

References 48 publications

(63 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It was reported that the resulting MLR model gave better results than the full data set. Egan et al (2001) presented some improvements in the computational neural network modeling process and pruning methods. They evaluated the performance of modified single value decomposition and optimal brain surgeon (OBS) pruning methods as proposed by Hassibi and Stork (1993).…”

Section: Artificial Neural Network-based Methods For Feature Selectionmentioning

confidence: 99%

Data Mining for Molecules with 2-D Neural Network Sensitivity Analysis

Embrechts

Arciniegas

Özdemir

et al. 2003

International Journal of Smart Engineering System Design

View full text Add to dashboard Cite

This paper illustrates a data mining application using two-dimensional (2-D) neural network sensitivity analysis for gaining insight into data strip mining problems. Data strip mining refers to predictive data mining problems where there are a large number of descriptive features, and the number of features is on the order of or exceeds the number of data records (e.g., 100 to 1000 features for 50 to 300 data records). After reducing the number of descriptive features to a manageable set using 1-D neural network sensitivity analysis (e.g., 40 features), a 2-D neural network sensitivity analysis allows the user to visualize variations in the response to identify relevant combinations of features. Each relevant combination can then be analyzed independently to look for interesting patterns and relationships, and can be used in this way to either prune more features or to get insight into the underlying rules for the model. 2-D sensitivity analysis enables the exploration of relevant relationships and features resulting in more robust, meaningful, and ef¢cient models. This methodology was applied to an in-silico drug design problem with 64 molecules and 160 d escriptive features.

show abstract

Section: Artificial Neural Network-based Methods For Feature Selectionmentioning

confidence: 99%

Data Mining for Molecules with 2-D Neural Network Sensitivity Analysis

Embrechts

Arciniegas

Özdemir

et al. 2003

International Journal of Smart Engineering System Design

View full text Add to dashboard Cite

show abstract

“…If the Hessian matrix is not definitely positive the minimum cannot be found. Inversion of the approximated Hessian matrix can be stabilized by using the LM algorithm [9,10]: (5) where I is the identity matrix, λ is the non-negative scalar and can be adjusted to maintain matrix H at the full rank.…”

Section: The Levenberg-marquardt Algorithmmentioning

confidence: 99%

“…However, the BP-MLFN method is slow, has temporal instability in training, and apt to fall into local minima. The Levenberg-Marquardt (LM) [9,10] algorithm has been used to modify the BP-MLFN method. The LM-BP-MLFN method is faster in convergence and robust.…”

Section: Introductionmentioning

confidence: 99%

Wavelet packet transform and artificial neural network applied to simultaneous kinetic multicomponent determination

Ren

Gao

2004

Analytical and Bioanalytical Chemistry

View full text Add to dashboard Cite

This paper presents a novel method, named wavelet packet transform based multilayer feedforward neural network with Levenberg-Marquardt and back propagation algorithm (WPTLMBP), developed for simultaneous kinetic determination of Cu(II), Fe(III), and Ni(II). Wavelet packet representations of signals provided a local time-frequency description, thus in the wavelet packet domain the quality of noise removal can be improved. The artificial neural network was applied for non-linear multivariate calibration. In this study, by optimization, wavelet packet function, decomposition level and number of hidden nodes for WPTLMBP method were selected as Db2, 2, and 4 respectively. A program PWPTLMBP was designed to perform simultaneous kinetic determination of Cu(II), Fe(III), and Ni(II). The relative standard error of prediction (RSEP) for all components with WPTLMBP, LM-BP-MLFN, and PLS methods were 6.39, 10.4, and 8.30%, respectively. Experimental results showed the proposed method to be successful and better than the others.

show abstract

“…On the other hand, ANNs have been successfully used in a wide range of applications in chemical processes modeling such as studying the quantitative structure‐activity relation of anti‐HIV activity for a large group of HEPT derivatives, predicting the binding affinities of thrombin and trypsin inhibitors, ion chromatography, and quantitating composition of mixtures of chlorinated hydrocarbons based on their Raman spectroscopy . However, building an ANN that gains success for a real world application is a very challenging problem even for highly experienced practitioners.…”

Section: Introductionmentioning

confidence: 99%

A multiobjective approach in constructing a predictive model for Fischer‐Tropsch synthesis

Dehghanian

Gheshlaghi

2017

Journal of Chemometrics

View full text Add to dashboard Cite

Fischer-Tropsch synthesis (FTS) is an important chemical process that produces a wide range of hydrocarbons. The exact mechanism of FTS is not yet fully understood, so prediction of the FTS products distribution is a not a trivial task. So far, artificial neural network (ANN) has been successfully applied for modeling varieties of chemical processes whenever sufficient and well-distributed training patterns are available. However, for most chemical processes such as FTS, acquiring such amount of data is very time-consuming and expensive. In such cases, neural network ensemble (NNE) has shown a significant generalization ability. An NNE is a set of diverse and accurate ANNs trained for the same task, and its output is a combination of outputs of these ANNs. This paper proposes a new NNE approach called NNE-NSGA-II that tries to prune this set by a modified nondominated sorting genetic algorithm to achieve an optimum subset according to 2 conflicting objectives, which are minimizing root-mean-square error in training and unseen data sets. Finally, a comparative study is performed on a single best ANN, a regular NNE, NNE-NSGA, and 3 popular ensemble of decision trees called random forest, stochastic gradient boosting, and AdaBoost.R2. The results show that in training data set, stochastic gradient boosting and AdaBoost.R2 have better fitted the samples; however, for the predicted FTS products in unseen data set, NNEs methods specially NNE-NSGA-II have considerably improved the generalization ability in comparison with the other competing approaches.

show abstract

Rapid optimization and minimal complexity in computational neural network multivariate calibration of chlorinated hydrocarbons using Raman spectroscopy

Cited by 9 publications

References 48 publications

Data Mining for Molecules with 2-D Neural Network Sensitivity Analysis

Data Mining for Molecules with 2-D Neural Network Sensitivity Analysis

Wavelet packet transform and artificial neural network applied to simultaneous kinetic multicomponent determination

A multiobjective approach in constructing a predictive model for Fischer‐Tropsch synthesis

Contact Info

Product

Resources

About