• Proceedings of the Korean Society of Near Infrared Spectroscopy Conference
• /
• 2001.06a
• /
• pp.1244-1244
• /
• 2001

In linear discriminant analysis there are two important properties concerning the effectiveness of discriminant function modeling. The first is the separability of the discriminant function for different classes. The separability reaches its optimum by maximizing the ratio of between-class to within-class variance. The second is the stability of the discriminant function against noises present in the measurement variables. One can optimize the stability by exploring the discriminant variates in a principal variation subspace, i. e., the directions that account for a majority of the total variation of the data. An unstable discriminant function will exhibit inflated variance in the prediction of future unclassified objects, exposed to a significantly increased risk of erroneous prediction. Therefore, an ideal discriminant function should not only separate different classes with a minimum misclassification rate for the training set, but also possess a good stability such that the prediction variance for unclassified objects can be as small as possible. In other words, an optimal classifier should find a balance between the separability and the stability. This is of special significance for multivariate spectroscopy-based classification where multicollinearity always leads to discriminant directions located in low-spread subspaces. A new regularized discriminant analysis technique, the principal discriminant variate (PDV) method, has been developed for handling effectively multicollinear data commonly encountered in multivariate spectroscopy-based classification. The motivation behind this method is to seek a sequence of discriminant directions that not only optimize the separability between different classes, but also account for a maximized variation present in the data. Three different formulations for the PDV methods are suggested, and an effective computing procedure is proposed for a PDV method. Near-infrared (NIR) spectra of blood plasma samples from mastitic and healthy cows have been used to evaluate the behavior of the PDV method in comparison with principal component analysis (PCA), discriminant partial least squares (DPLS), soft independent modeling of class analogies (SIMCA) and Fisher linear discriminant analysis (FLDA). Results obtained demonstrate that the PDV method exhibits improved stability in prediction without significant loss of separability. The NIR spectra of blood plasma samples from mastitic and healthy cows are clearly discriminated between by the PDV method. Moreover, the proposed method provides superior performance to PCA, DPLS, SIMCA and FLDA, indicating that PDV is a promising tool in discriminant analysis of spectra-characterized samples with only small compositional difference, thereby providing a useful means for spectroscopy-based clinic applications.

• Proceedings of the Korean Society of Near Infrared Spectroscopy Conference
• /
• 2001.06a
• /
• pp.1042-1042
• /
• 2001

In linear discriminant analysis there are two important properties concerning the effectiveness of discriminant function modeling. The first is the separability of the discriminant function for different classes. The separability reaches its optimum by maximizing the ratio of between-class to within-class variance. The second is the stability of the discriminant function against noises present in the measurement variables. One can optimize the stability by exploring the discriminant variates in a principal variation subspace, i. e., the directions that account for a majority of the total variation of the data. An unstable discriminant function will exhibit inflated variance in the prediction of future unclassified objects, exposed to a significantly increased risk of erroneous prediction. Therefore, an ideal discriminant function should not only separate different classes with a minimum misclassification rate for the training set, but also possess a good stability such that the prediction variance for unclassified objects can be as small as possible. In other words, an optimal classifier should find a balance between the separability and the stability. This is of special significance for multivariate spectroscopy-based classification where multicollinearity always leads to discriminant directions located in low-spread subspaces. A new regularized discriminant analysis technique, the principal discriminant variate (PDV) method, has been developed for handling effectively multicollinear data commonly encountered in multivariate spectroscopy-based classification. The motivation behind this method is to seek a sequence of discriminant directions that not only optimize the separability between different classes, but also account for a maximized variation present in the data. Three different formulations for the PDV methods are suggested, and an effective computing procedure is proposed for a PDV method. Near-infrared (NIR) spectra of blood plasma samples from daily monitoring of two Japanese cows have been used to evaluate the behavior of the PDV method in comparison with principal component analysis (PCA), discriminant partial least squares (DPLS), soft independent modeling of class analogies (SIMCA) and Fisher linear discriminant analysis (FLDA). Results obtained demonstrate that the PDV method exhibits improved stability in prediction without significant loss of separability. The NIR spectra of blood plasma samples from two cows are clearly discriminated between by the PDV method. Moreover, the proposed method provides superior performance to PCA, DPLS, SIMCA md FLDA, indicating that PDV is a promising tool in discriminant analysis of spectra-characterized samples with only small compositional difference.

• Proceedings of the Korean Society of Near Infrared Spectroscopy Conference
• /
• 2001.06a
• /
• pp.1256-1256
• /
• 2001

The physico-chemical and texture characteristics of meat determine the nutritional, technological and sensory quality. However, the analysis of meat quality requires expensive, laborious and time consuming analytical methods. The objective of this study was to evaluate NIR spectroscopy using transmittance for determining the moisture, fat, protein and total pigment content, the water holding capacity (WHC) and the toughness of beef meat. A total of 318 spectra were recorded from ground beef samples by a Feed Analyzer 1265 of Infratec. The samples were obtained from the Longissimus muscle of the 10$^{th}$ rib of yearling bulls, ground with an electrical chopper, vacuum packaged, aged during 7 days and frozen at -24$^{\circ}C$ until the analyses were done. Moisture content was measured by oven drying at 10$0^{\circ}C$, fat content was determined by Soxhlet extraction and protein content was estimated from nitrogen content using the Kjeldahl analysis. The total pigment content was determined by the method of Hornsey and the WHC using the method of filter paper press. The instrumental evaluation of texture (maximum load WB, maximum stress MS and toughness) was conducted in an Instron equipment with a Warner-Bratzler shearing device. This analysis was performed on a chop of 3.5 cm obtained from the longissimus of the 8$^{th}$ rib, aged during 7 days, kept frozen at -24$^{\circ}C$ and cooked before the analysis. Near infrared spectra were recorded as log 1/T (T=transmittance) at 2 nm intervals from 850 to 1050 nm using a Feed Analyzer 1265 of Infratec. Calibrations were performed with the WinISI software (vs. 1.02) using the MPLS method. To examine the effect of scatter correction o. derivation of spectra on the calibration performance, calibrations were calculated with the crude spectra or pretreated with different mathematical treatments (inverse MSC, SNVD) and/or second derivative operation. For chemical composition, the use of the scatter corrections improved the calibration statistics, in terms of lower SECV and higher $r^2$. In most of the variables, the use of the 2$^{nd}$ derivative improved the predictions, mainly when combined with the SNVD treatment. However, for predicting the texture traits, the best estimation was obtained from the crude spectrum. These results showed that the equations obtained for predicting moisture, fat and total pigments were very accurate, with $r^2$ being higher that 0.9. However, the prediction of the texture traits (WB, MS, toughness) from ground meat was poor.

• Journal of Information Display
• /
• v.2 no.3
• /
• pp.18-31
• /
• 2001

We utilize Fourier transform infrared (IR) spectroscopy to probe intramolecular hydrogen bonding in $smectic-C^{\ast}$ liquid crystal phases. Infrared spectra of aligned smectic liquid crystal materials vs. temperature and of isotropic liquid crystal mixtures vs. concentration were measured in homologs, both with and without hydrogen bonding. Hydrogen bonding significantly changes the direction and magnitude of the vibrational dipole transition moments, causing marked changes in the IR dichroic absorbance profiles of hydrogen bonded molecular subfragments. A GAUSSIAN94 computation of the directions, magnitudes, and frequencies of the vibrational dipole moments of molecular subfragments shows good agreement with the experimental data. The results show that IR dichroism can be an effective probe of hydrogen bonding in liquid crystal phases.

• Near Infrared Analysis
• /
• v.1 no.1
• /
• pp.9-17
• /
• 2000

A variety of standardization methods between two near-infrared (NIR) spectrometers were investigated for the prediction of five constituents in trans-alkylation process. Spectra were collected by two different instruments (one is regarded as mater instrument, other on as slave instrument). Three well-known standardization methods of direct standardization (DS), piecewise direct standardization (PDS) and slope/bias correction of response variable were applied to trans-alkylation samples on the slave instrument. We have examined for a set of reliable standardization samples using smaller number of transfer samples in order to increase efficiency of standardization.

• Near Infrared Analysis
• /
• v.1 no.1
• /
• pp.23-26
• /
• 2000

The objectives of this study were to examine the ability to predict soluble solid and firmness in intact apple based on the visible/near-infrared spectroscopic technique. Two cultivars of apples, Delicious and Gala, were handled, tested and analyzed. Reflectance spectra, Magness-Taylor (MT) Firmness, and soluble solids in apples were measured sequentially. Maximum and minimum diameters, height, and weight of apples were recorded before the MT firmness tests. Apple samples were divided in to a calibration set and a validation set. The method of partial least squares (PLS) analysis was used. a unique set of PLS loading vectors (factors) was development for soluble solid and firmness. The PLS model showed good relationship between predicted and measured soluble solids in intact apples in the wavelength range of 860∼1078 nm. However, the PLS analysis was not good enough to predict the apple firmness.

• Proceedings of the Korean Society of Near Infrared Spectroscopy Conference
• /
• 2001.06a
• /
• pp.2102-2102
• /
• 2001

Near infrared, diffuse reflectance spectroscopy has shown significant potential for in vitro and in vivo assessment of metabolic status. However, the complexity of living samples can lead to ambiguous results. This presentation will focus on methods that provide controls for scattering and absorption estimation in tissue. For robust estimations, normalization procedures will be shown which can greatly improve interpretability of results. Normalization based on time, location and spectral property will be shown with data from models, tissue phantoms and in vivo measurements. In particular, interpretation of NIR spectra associated with major respiratory constituents will be examined. Measurement of constituents such as hemoglobin, myoglobin, tissue edema, and lactate will be shown. Results suggest that NIR may provide a valuable tool for physiological monitoring in critical care research and practice.

• Proceedings of the Korean Society of Near Infrared Spectroscopy Conference
• /
• 2001.06a
• /
• pp.1628-1628
• /
• 2001

Near infrared (NIR) spectroscopy is a powerful technique for non-destructive analysis that can be obtained in a wide range of environments. Recently, NIR measurements have been utilized as probe for quantitative analysis in agricultural, industrial, and medical sciences. In addition, it is also possible to make practical application on NIR for molecular structural analysis. In this work, Fourier transform near infrared (FT-NIR) measurements were carried out to utilize extensively in the relative amounts of different secondary structures were employed, such as Iysozyme, concanavalin A, silk fibroin and so on. Several broad NIR bands due to the protein absorption were observed between 4000 and $5000\;^{-1}$. In order to obtain more structural information from these featureless bands, second derivative and Fourier-self-deconvolution procedures were performed. Significant band separation was observed near the feature at $4610\;^{-1}$ ,. Particularly the peak intensity at $4525\;^{-1}$ shows a characteristic change with thermal denaturation of fibroin. The structural information can be also obtained by mid-IR and CD spectral. Correlation of NIR spectra with protein structure is discussed.

• Proceedings of the Korean Society of Near Infrared Spectroscopy Conference
• /
• 2001.06a
• /
• pp.1514-1514
• /
• 2001

In this paper a fast non-destructive analytical method to measure various nutrients in the intact tomato---Near infrared Spectrometry NIRs was introduced Using this method the content of some organic acid, vitamin C, reductive sugar, and solid soluble were determined simultaneously. Screen out four wavelengths at 916nm, 1000nm, 1004nm and 832nm to present optimum four optical terms of d$^2$ log(1/R) with second derivative spectra treating data scanned under these wavelengths. The multiple correlation coefficients between these values and those obtained on chemical analysis were 0.983, 0.990, 0.987, and 0.994, respectively, and the standard errors of prediction (SEP) were 0.007, 0.440, 0.037, and 0.057, respectively. These results indicate that NIRs is comparable to chemical methods in both accuracy and precision and is reliable method for determination of nutrients in intact tomato.

• YAKHAK HOEJI
• /
• v.47 no.5
• /
• pp.261-265
• /
• 2003

Near-infrared (NIR) system was used to determine rapidly and simply indomethacin in buffer solution for a dissolution test of tablets and capsules. Indomethacin standards were prepared ranging from 10 to 50 ppm using the mixture of phosphate buffer (pH 7.2) and water (1 : 4). The near-infrared (NIR) transmittance spectra of indomethacin standard solutions were collected by using a quartz cell in 1 mm and 2 mm pathlength. Partial least square regression (PLSR) was explored to develop calibration models over the spectral range 1100∼1700 nm. The model using 1 mm quartz cell was better than that using 2 mm quartz cell. The PLSR models developed gave standard error of prediction (SEP) of 0.858 ppm. In order to validate the developed calibration model, routine analysis was performed using another standard solutions. The NIR routine analysis showed good correlation with actual values. Standard error of prediction (SEP) is 1.414 ppm for 7 indomethacin samples in routine analysis and its error was permeable in the regulation of Korean Pharmacopoeia (VII). These results show the potential use of the real time monitoring for indomethacin during a dissolution test.