Optimal minimal neural interpretation of spectra

Borggaard, Claus; Thodberg, Hans Henrik

doi:10.1021/ac00029a018

Cited by 264 publications

(125 citation statements)

References 7 publications

Supporting

Mentioning

124

Contrasting

Order By: Relevance

“…Table 7, originally taken from Eilers et al (2009) but modified to include our results, show the SEP results for neural networks (see Borggaard and Thodberg, 1992;Thodberg, 1995). For instance, our method is roughly 3.5 times more efficient than the 13-X-1 network in the prediction sense.…”

Section: Multiple Linear Regression But Have Lower Accuracy and Precimentioning

confidence: 99%

See 1 more Smart Citation

Aggregated functional data model for near-infrared spectroscopy calibration and prediction

Dias

Garcia

Ludwig

et al. 2014

Journal of Applied Statistics

View full text Add to dashboard Cite

Calibration and prediction for NIR spectroscopy data are performed based on a functional interpretation of the Beer-Lambert formula. Considering that, for each chemical sample, the resulting spectrum is a continuous curve obtained as the summation of overlapped absorption spectra from each analyte plus a Gaussian error, we assume that each individual spectrum can be expanded as a linear combination of B-splines basis. Calibration is then performed using two procedures for estimating the individual analytes curves: basis smoothing and smoothing splines.Prediction is done by minimizing the square error of prediction. To assess the variance of the predicted values, we use a leave-one-out jackknife technique. Departures from the standard error models are discussed through a simulation study, in particular, how correlated errors impact on the calibration step and consequently on the analytes' concentration prediction. Finally, the performance of our methodology is demonstrated through the analysis of two publicly available datasets.

show abstract

Section: Multiple Linear Regression But Have Lower Accuracy and Precimentioning

confidence: 99%

“…Borggaard and Thodberg (1992) cite the correlation as an important issue in the analysis of spectra, since the fine sampling usually results in large correlation between adjacent points in the spectrum. Also, we apply the non-parametric analysis to two real datasets.…”

Section: Introductionmentioning

confidence: 99%

Aggregated functional data model for near-infrared spectroscopy calibration and prediction

Dias

Garcia

Ludwig

et al. 2014

Journal of Applied Statistics

View full text Add to dashboard Cite

show abstract

“…The multilayer feed-forward neural network trained with back-propagation learning algorithm becomes an increasingly popular technique. [23][24][25] Recently, we reported the application of ANN for non-linear calibration by using potentiometric titration and spectrophotometry. [26][27][28] In this study, three chemometric methods were successfully applied to simultaneous determination of PCT and CAF in a commercial tablet formulation, tablets without any separation procedure.…”

Section: Introductionmentioning

confidence: 99%

Spectrophotometric Simultaneous Determination of Caffeine and Paracetamol in Commercial Pharmaceutical by Principal Component Regression, Partial Least Squares and Artificial Neural Networks Chemometric Methods

Aktaş

Kitiş

2014

Croat. Chem. Acta

View full text Add to dashboard Cite

Abstract. Three multivariate calibration-prediction techniques, principal component regression (PCR), partial least squares (PLS) and artificial neural networks (ANN) were applied to the spectrometric multicomponent analysis of the drug containing paracetamol (PCT) and caffeine (CAF) without any separation step. The selection of variables was studied. A series of synthetic solution containing different concentrations of PCT and CAF were used to check the prediction ability of the PCR, PLS and ANN. The results obtained in this investigation strongly encourage us to apply these techniques for a routine analysis and quality control of the drug.(doi: 10.5562/cca2214)

show abstract

“…As powerful analytical tools, spectroscopic techniques in combination with calibration models have seen increasing implementation in sectors as diverse as food, pharmaceuticals and petrochemical [1][2][3]. However, as a consequence of the large number of spectral wavelengths, multivariate calibration methods, such as partial least squares (PLS) [4] [5] and other approaches including neural networks [1] [6] and Gaussian process models [7], are required for the development of robust calibration models.…”

Section: Introductionmentioning

confidence: 99%

“…However, as a consequence of the large number of spectral wavelengths, multivariate calibration methods, such as partial least squares (PLS) [4] [5] and other approaches including neural networks [1] [6] and Gaussian process models [7], are required for the development of robust calibration models. One challenge faced in industrial on-line and in-line applications of spectroscopy is that the samples are not analyzed under well-controlled laboratory conditions, materializing in fluctuations in some of the external factors, such as temperature and pressure.…”

Section: Introductionmentioning

confidence: 99%

The impact of temperature variations on spectroscopic calibration modelling: a comparative study

Chen

Martin

2007

Journal of Chemometrics

View full text Add to dashboard Cite

Abstract:Temperature fluctuations can have a significant impact on the repeatability of spectral measurements and as a consequence can adversely affect the resulting calibration model. More specifically, when test samples measured at temperatures unseen in the training data set are presented to the model, degraded predictive performance can materialize. Current methods for addressing the temperature variations in a calibration model can be categorized into two classescalibration model based approaches, and spectra standardization methodologies. This paper presents a comparative study on a number of strategies reported in the literature including partial least squares (PLS), continuous piecewise direct standardization (CPDS) and loading space standardization (LSS), in terms of the practical applicability of the algorithms, their implementation complexity, and their predictive performance. It was observed from the study that the global modelling approach, where latent variables are initially extracted from the spectra using PLS, and then augmented with temperature as the independent variable, achieved the best predictive performance. In addition, the two spectra standardization methods, CPDS and LSS, did not provide consistently enhanced performance over the conventional global modelling approach, despite the additional effort in terms of standardizing the spectra across different temperatures. Considering the algorithmic complexity and resulting calibration accuracy, it is concluded that the global modelling (with temperature) approach should be first considered for the development of a calibration model where temperature variations are known to affect the fundamental data, prior to investigating the more powerful spectra standardization approaches.

show abstract

Optimal minimal neural interpretation of spectra

Cited by 264 publications

References 7 publications

Aggregated functional data model for near-infrared spectroscopy calibration and prediction

Aggregated functional data model for near-infrared spectroscopy calibration and prediction

Spectrophotometric Simultaneous Determination of Caffeine and Paracetamol in Commercial Pharmaceutical by Principal Component Regression, Partial Least Squares and Artificial Neural Networks Chemometric Methods

The impact of temperature variations on spectroscopic calibration modelling: a comparative study

Contact Info

Product

Resources

About