SEARCH · Search PubMed

Results for “Discriminant Analysis”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 955 records · Page 53Linked to original sources

PubMed2005

Determination of total nitrogen content, pH, density, refractive index, and brix in Thai fish sauces and their classification by near-infrared spectroscopy with searching combination moving window partial least squares.

Near-infrared (NIR) transflectance spectra in the region of 1100-2500 nm were measured for 100 Thai fish sauces. Quantitative analyses of total nitrogen (TN) content, pH, refractive index, density and brix in the Thai fish sauces and their qualitative analyses were carried out by multivariate analyses with the aid of wavelength interval selection method named searching combination moving window partial least squares (SCMWPLS). The optimized informative region for TN selected by SCMWPLS was the region of 2264-2428 nm. A PLS calibration model, which used this region, yielded the lowest root mean square error of prediction (RMSEP) of 0.100% w/v for the PLS factor of 5. This prediction result is significantly better than those obtained by using the whole spectral region or informative regions selected by moving window partial least squares regression (MWPLSR). As for pH, density, refractive index and brix, the 1698-1722, and 2222-2258 nm regions, the 1358-1438 nm region, the 1774-1846, and 2078-2114 nm regions, and the 1322-1442, and 2000-2076 nm regions were selected by SCMWPLS as the optimized regions. The best prediction results were always obtained by use of the optimized regions selected by SCMWPLS. The lowest RMSEP for pH, density, refractive index and brix were 0.170, 0.007 g cm(-3), 0.0079 and 0.435 degrees Brix, respectively. Qualitative models were developed by using four supervised pattern recognitions, linear discriminant analysis (LDA), factor analysis-linear discriminant analysis (FA-LDA), soft independent modeling of class analog (SIMCA), and K neareat neighbors (KNN) for the optimized combination of informative regions of the NIR spectra of fish sauces to classify fish sauces into three groups based on TN. All the developed models can potentially classify the fish sauces with the correct classification rate of more than 82%, and the KNN classified model has the highest correct classification rate (95%). The present study has demonstrated that NIR spectroscopy combined with SCMWPLS is powerful for both the quantitative and qualitative analyses of Thai fish sauces.

Animals↗

PubMed2003

Comparison of statistical methods for classification of ovarian cancer using mass spectrometry data.

MOTIVATION: Novel methods, both molecular and statistical, are urgently needed to take advantage of recent advances in biotechnology and the human genome project for disease diagnosis and prognosis. Mass spectrometry (MS) holds great promise for biomarker identification and genome-wide protein profiling. It has been demonstrated in the literature that biomarkers can be identified to distinguish normal individuals from cancer patients using MS data. Such progress is especially exciting for the detection of early-stage ovarian cancer patients. Although various statistical methods have been utilized to identify biomarkers from MS data, there has been no systematic comparison among these approaches in their relative ability to analyze MS data. RESULTS: We compare the performance of several classes of statistical methods for the classification of cancer based on MS spectra. These methods include: linear discriminant analysis, quadratic discriminant analysis, k-nearest neighbor classifier, bagging and boosting classification trees, support vector machine, and random forest (RF). The methods are applied to ovarian cancer and control serum samples from the National Ovarian Cancer Early Detection Program clinic at Northwestern University Hospital. We found that RF outperforms other methods in the analysis of MS data.

Algorithms↗

PubMed2006

Differentiation of faecal Escherichia coli from humans and animals by multiple antibiotic resistance analysis.

AIMS: Multiple antibiotic resistance (MAR) was performed on 128 Escherichia coli isolates, recovered from faecal samples of humans and animals (cattle, goat, sheep) to determine and compare their antibiotic resistance patterns and to evaluate them statistically in order to specify the source of the faecal material. METHODS AND RESULTS: Disk diffusion method was applied with a selection of antibiotics. Statistical approach was performed with hierarchical cluster analysis (CA), discriminant analysis (DA) and principal component analysis. Comparing human and animal isolates there was significant difference in levels of resistance to all antibiotics tested (P<0.05) with 46 and 24 distinct resistance patterns for human and animal isolates respectively. CA and DA separated human and animal isolates with a high average rate of correct classification (99.2%), when all animal isolates were pooled together. CONCLUSIONS: MAR analysis compared with appropriate statistical evaluation may provide a useful tool for differentiating the human or animal origin of E. coli isolates derived from environmental samples. Subsequently, determination of the source of faecal pollution becomes possible. SIGNIFICANCE AND IMPACT OF THE STUDY: Determining the source of faecal pollution enables the prediction of possible risk for public health and the application of appropriate management plans for prevention of further contamination.

Animals↗

PubMed1990

[A comparison of hematological and serum biochemical values between two groups of female cynomolgus monkeys reared under different conditions].

The influence of long term rearing conditions on the hematological and serum biochemical values in laboratory-bred female cynomolgus monkeys (Macaca fascicularis) was studied under the following two conditions. Group A: New born infants were nursed by their mothers for at least 12 weeks after birth. After weaning, four monkeys of approximately equal age and body weight were kept in a relatively small cage (90 cm W x 60 cm D x 60 cm H) and reared to five years of age. These rearing conditions are commonly used in the breeding program of our primate center. Group B: One male and five females of wild origin were introduced into a relatively large cage (180 cm W x 180 cm D x 160 cm H). Female monkeys born in these cages were used in this study. All of them were reared by their own mothers and grew well in these cages. Male monkeys born in these cages were removed at about 3.5 years of age. Hematological and serum biochemical values between the two groups in each age class were compared and analyzed by the canonical discriminant analysis (the discriminant analysis with reduction of dimensionality). Age correlated highly with the value of the first canonical variate. The second canonical variate detected a difference in the blood hematology and biochemistry probably induced by the difference in rearing conditions.

Age Factors↗

PubMed2005

Cephalometric predictors of long-term stability in the early treatment of Class III malocclusion.

The aim of this study was to examine the differences in the early craniofacial morphology of Class Ill malocclusions. Lateral cephalograms of 45 subjects with a Class III malocclusion and an anterior crossbite in the deciduous or mixed dentition were examined before treatment, after treatment, and during the long-term retention stage. The anterior crossbites of all patients were corrected after a series of orthodontic treatments. After a mean follow-up period of 5.7 years, all the subjects were reevaluated and divided into three groups according to the final occlusal status: good, fair, and poor occlusal stability. Twenty cephalometric variables on the pretreatment lateral cephalograms were analyzed by one-way analysis of variance and discriminant analysis to identify the key determinants for discriminating among the three groups. Among the 20 variables, 11 showed statistical significance. Generally, the subjects with a smaller gonial angle and a more hypodivergent skeletal pattern had good prognosis after the early treatment of Class III malocclusion. When the AB to mandibular plane angle and N-perpendicular to point A were selected in discriminant analysis, the AB to mandibular plane angle was the most significant variable. Discriminant analysis showed a relatively high degree of correct classifications of the patients with early Class III malocclusion. In particular, discriminant analysis showed the highest accuracy (93.3%) when predicting a poor prognosis.

Adolescent↗

PubMed2006

Use of unsupervised and supervised artificial neural networks for the identification of lactic acid bacteria on the basis of SDS-PAGE patterns of whole cell proteins.

Conventional multivariate statistical techniques (hierarchical cluster analysis, linear discriminant analysis) and unsupervised (Kohonen Self Organizing Map) and supervised (Bayesian network) artificial neural networks were compared for as tools for the classification and identification of 352 SDS-PAGE patterns of whole cell proteins of lactic acid bacteria belonging to 22 species of the genera Lactobacillus, Leuconostoc, Enterococcus, Lactococcus and Streptococcus including 47 reference strains. Electrophoretic data were pre-treated using the logistic weighting function described by Piraino et al. [Piraino, P., Ricciardi, A., Lanorte, M. T., Malkhazova, I., Parente, E., 2002. A new procedure for data reduction in electrophoretic fingerprints of whole-cell proteins. Biotechnol. Lett. 24, 1477-1482]. Hierarchical cluster analysis provided a satisfactory classification of the patterns but was unable to discriminate some species (Leuconostoc, Lb. sakei/Lb. curvatus, Lb. acidophilus/Lb. helveticus, Lb. plantarum/Lb. paraplantarum, Lc. lactis/Lc. raffinolactis). A 7x7 Kohonen self-organizing map (KSOM), trained with the patterns of the reference strains, provided a satisfactory classification of the patterns and was able to discriminate more species than hierarchical cluster analysis. The map was used in predictive mode to identify unknown strains and provided results which in 85.5% of cases matched the classification obtained by hierarchical cluster analysis. Two supervised tools, linear discriminant analysis and a 23:5:2 Bayesian network were proven to be highly effective in the discrimination of SDS-PAGE patterns of Lc. lactis from those of other species. We conclude that data reduction by logistic weighting coupled to traditional multivariate statistical analysis or artificial neural networks provide an effective tool for the classification and identification of lactic acid bacteria on the basis of SDS-PAGE patterns of whole cell proteins.

Bacterial Proteins↗

PubMed1996

Application of artificial neural networks for the classification of liver lesions by image texture parameters.

Ultrasound imaging is a powerful tool for characterizing the state of soft tissues; however, in some cases, where only subtle differences in images are seen as in certain liver lesions such as hemangioma and malignancy, existing B-scan methods are inadequate. More detailed analyses of image texture parameters along with artificial neural networks can be utilized to enhance differentiation. From B-scan ultrasound images, 11 texture parameters comprising of first, second and run length statistics have been obtained for normal, hemangioma and malignant livers. Tissue characterization was then performed using a multilayered backpropagation neural network. The results for 113 cases have been compared with a classification based on discriminant analysis. For linear discriminant analysis, classification accuracy is 79.6% and with neural networks the accuracy is 100%. The present results show that neural networks classify better than discriminant analysis, demonstrating a much potential for clinical application.

Diagnosis, Differential↗

PubMed1989

Disturbed endocrine function in the psychoses. II: Discriminant function analysis of multihormone data.

Discriminant function analyses were performed on plasma concentrations of prolactin, growth hormone, cortisol, TSH, and the neurophysins measured over 17 hours in 70 newly admitted drug-free psychiatric patients and 35 control subjects. The hormone data distinguished between patients with different classes of drug-free psychosis (26 schizophrenic, 12 with manic disorder, 10 with major depressive disorder, psychotic subtype, 9 with schizoaffective mania (SAM]. Higher plasma cortisol and lower TSH concentrations separated eight of nine SAM patients from all others.

Bipolar Disorder↗

PubMed1992

Prediction of severe coronary artery disease using computerized ECG measurements and discriminant function analysis.

This study tested the hypothesis that discriminant function analysis of clinical and exercise-test variables including computerized ST measurements could improve the prediction of severe coronary artery disease. Secondary objectives were to demonstrate the effect of digoxin and/or resting electrocardiographic (ECG) abnormalities, and to evaluate the relative importance of ST measurements made during the recovery phase and in the three lead group areas. The design was a retrospective analysis of data collected during exercise testing and coronary angiography. The ECG data were gathered and stored in digital format on optical discs and all ST measurements were made off-line using the authors' own software. Univariate and multivariate analytic methods were used to analyze all pretest characteristics as well as hemodynamic and computerized ECG responses to exercise. A 1,000-bed Veterans Affairs Medical Center served as the setting. The study included 446 male veterans who underwent a sign or symptom limited treadmill exercise test and coronary angiography. Analysis was also performed on a subset of this population formed by excluding patients receiving digoxin or with resting ECGs exhibiting left ventricular hypertrophy or ST depression (n = 328). In the total study population, the authors derived a treadmill score using discriminant function analysis. This score included: (1) the time-slope area in lead V5 during recovery; (2) delta heart rate; (3) angina pectoris during the exercise test; and (4) presence of diagnostic Q waves on the resting ECG. This score was effective in predicting triple vessel/left main disease and outperformed exercise-induced ST depression for predicting severe coronary artery disease. After exclusion of patients with ECGs exhibiting left ventricular hypertrophy or resting ST depression and patients receiving digoxin, discriminant function analysis chose: (1) the time-slope area in lead V5 during recovery and (2) delta heart rate. Exclusion of these patients resulted in a nonsignificant decrease in specificity of all ST criteria. ST-segment amplitude or slope in lead V5 at 3.5 minutes in recovery clearly outperformed the maximal exercise measurements in both groups. Summing the depressions or selecting the most depression in the three areas (ie, lateral-V5, inferior-II, anterior-V2) did not improve test performance. Leads other than V5 did not contain significant diagnostic information. A quantitative approach to exercise testing using discriminant function analysis enhanced the tests' performance for predicting severe coronary disease. The inclusion of patients taking digoxin or with resting ECG abnormalities nonsignificantly decreases the specificity of all ST criteria.(ABSTRACT TRUNCATED AT 400 WORDS)

Adult↗

PubMed2000

Discrimination and identification of coastal Douglas-fir clones using needle flavonoid fingerprints.

This paper describes a method for discriminating and identifying 10 successful Douglas-fir (Pseudotsuga menziesii var. menziesii) clones using foliar flavonoids. All the 101 individuals analyzed by high performance liquid chromatography contained two proanthocyanidins: prodelphinidin and procyanidin and six flavonols: myricetin, quercetin, larycitrin, kaempferol, isorhamnetin and syringetin, but in different proportions. The experimental protocol used was very reproducible since the variation coefficients for each flavonoid did not exceed 9%. Submission of the flavonoid data to multivariate discriminant analysis allowed excellent discrimination of the 10 clones with 89% of the individuals being well-grouped. Then a clonal bank was established in which the fingerprint of each clone is defined by its position in the multidimensional space of the discriminant analysis. The clonal identity of several unknown individuals was determined with success by projecting their flavonoid data in a subsequent discriminant analysis.

Journal Article↗

PubMed1984

A discriminant-function analysis of screening tests for excessive drinking and alcoholism.

Because a previous study among 385 psychiatric admissions had shown each of three rapid interviews to be far superior to each of nine laboratory tests in screening for excessive drinking and alcoholism, the separation of patients with these drinking patterns from normal drinkers was reexamined by the more sophisticated technique of discriminant analysis. It was thus possible to determine where there was overlap in the information provided by some tests in contrast to "new information" provided by others and whether the arbitrary cut-off points of the normal ranges of the laboratory tests were contributing to their poor sensitivity. Discriminant analysis again confirmed the good performance of the rapid interviews, particularly the Brief Michigan Alcoholism Screening Test and the Reich interview, but it also identified glutamate dehydrogenase (GDH) as the best of the laboratory tests and of comparable efficacy to the rapid interview for the group of excessive drinkers. By comparison, gamma-glutamyl transpeptidase and mean corpuscular volume performed poorly. Using the whole range of results rather than a single cut-off point for discriminant analysis did not alter the relative performance of the screening tests. The optimum combination of tests was that of the Reich interview and the GDH, achieving 100% sensitivity for excessive drinking and alcoholism without any decline in the specificity or predictive value of a positive test result.

Adult↗

PubMed1981

Genetic relationships between sublines of inbred strains of mice as assessed by mandible analysis.

Genetic relationships between sublines of inbred strains of mice were investigated by using mandible measurements. (1) Canonical discriminant analysis (discriminant analysis with reduction of dimensionality) was applied to the investigation of genetic relationships between sublines on two kinds of dimensionality. As a result, it was made clear that all the sublines of the same strains mapped closer to each other than they were to other strains on the plane of z1--z2, except the sublines of strain AA. (2) According to the result of discriminant analysis, misclassification among the sublines of the other strains was only 0 head/302 head, or 0%, for the male mice and 6 head/302 head, or 1.99%, for the female mice. Misidentification between the sublines of the same strain was also of low rate; that is, 28/302, or 9.27%, or the males and 9/302, or 2.98%, for the females. These results suggested that the mice of the sublines of the same strain might be genetically different from one another. (3) The cause for the occurrence of such subline divergence as this seemed to be mainly due to the spontaneous mutation of each colony after the separation of mice. It appeared that the subline divergence of strain AA might be due to the genetic heterozygosity of an original stock before the separation of mice.

Animals↗

PubMed2005

KPCA plus LDA: a complete kernel Fisher discriminant framework for feature extraction and recognition.

This paper examines the theory of kernel Fisher discriminant analysis (KFD) in a Hilbert space and develops a two-phase KFD framework, i.e., kernel principal component analysis (KPCA) plus Fisher linear discriminant analysis (LDA). This framework provides novel insights into the nature of KFD. Based on this framework, the authors propose a complete kernel Fisher discriminant analysis (CKFD) algorithm. CKFD can be used to carry out discriminant analysis in "double discriminant subspaces." The fact that, it can make full use of two kinds of discriminant information, regular and irregular, makes CKFD a more powerful discriminator. The proposed algorithm was tested and evaluated using the FERET face database and the CENPARMI handwritten numeral database. The experimental results show that CKFD outperforms other KFD algorithms.

Algorithms↗

PubMed1989

[Age changes of functional nuclear swelling and classes of nuclei in human hepatocytes].

The aim of the study was to describe aging changes in the functional nuclear edema and the classification of human hepatocyte nuclei, by determining three parameters--the size of the nucleus, the relative DNA quantity and the number of chromocenters. For this purpose, karyometry and DNA cytophotometry were performed on 10 human liver preparations. The data obtained was subjected to correlation, cluster and discriminance analysis. The results indicated a reduction in the capacity of liver cells for functional nuclear edema as aging progressed. Whereas at a young age there is only a loose correlation between nuclear size and DNA content, it becomes much stricter later on, partly caused by polyploidization. Cluster analysis, followed by discriminance analysis, is well suited for dividing the nuclei of human hepatocytes into two or three statistical populations provided the nuclear area, DNA quantity and number of chromocenters are all used as characteristics simultaneously. When allowance is made for functional edema, the biological interpretation of clusters from young liver preparations permits meaningful conclusions, but it appears problematic for old preparations. Here it might be more practical to analyze the mixed distributions resulting from a determination of the DNA quantity or nuclear size alone.

Aging↗

PubMed2004

Classification of signatures of Bovine Spongiform Encephalopathy in serum using infrared spectroscopy.

Signatures of Bovine Spongiform Encephalopathy (BSE) have been identified in serum by means of "Diagnostic Pattern Recognition (DPR)". For DPR-analysis, mid-infrared spectroscopy of dried films of 641 serum samples was performed using disposable silicon sample carriers and a semi-automated DPR research system operating at room temperature. The combination of four mathematical classification approaches (principal component analysis plus linear discriminant analysis, robust linear discriminant analysis, artificial neural network, support vector machine) allowed for a reliable assignment of spectra to the class "BSE-positive" or "BSE-negative". An independent, blinded validation study was carried out on a second DPR research system at the Veterinary Laboratory Agency, Weybridge, UK. Out of 84 serum samples originating from terminally-ill, BSE-positive cattle, 78 were classified correctly. Similarly, 73 out of 76 BSE-negative samples were correctly identified by DPR such that, numerically, an accuracy of 94.4 % can be calculated. At a confidence level of 0.95 (alpha = 0.05) these results correspond to a sensitivity > 85% and a specificity > 90%. Identical class assignment by all four classifiers occurred in 75% of the cases while ambiguous results were obtained in only 8 of the 160 cases. With an area under the ROC (receiver operating charateristics) curve of 0.991, DPR may potentially supply a valuable surrogate marker for BSE even in cases in which a deliberate bias towards improved sensitivity or specificity is desired. To the best of our knowledge, DPR is the first and--up to now--only method which has demonstrated its capability of detecting BSE-related signatures in serum.

Animals↗

PubMed1995

Modified neonatal growth assessment score: a multivariate approach to the detection of intrauterine growth retardation in the neonate.

The objective of this investigation was the development of a modified Neonatal Growth Assessment Score (mNGAS) for use in the evaluation of neonatal growth status. The growth of 74 fetuses at risk for intrauterine growth retardation (IUGR) was followed longitudinally, beginning in the first or early second trimester. Rossavik growth models derived from data obtained in the second trimester were used to predict the weight (WT), crown-heel length (CHL) and head (HC), abdominal (AC) and thigh circumferences (ThC) at birth, which were then actually measured within 24 h after delivery. These measurements were compared to age-specific size curves and used to calculate sets of five growth potential index (GPRIi) values, which in turn were used to calculate five-variable Neonatal Growth Assessment Scores (NGAS5). Neonates were initially classified as normal or IUGR on the basis of NGAS5, GPRIi, and anatomic measurements. A final classification based on principal component analysis and linear discriminant analysis was carried out. The score obtained using the first principal component function was defined to be mNGAS51. The effectiveness of mNGASij values, determined from 1-4 GPRIi values, in separating normal and IUGR neonates was also evaluated. Neonates initially considered to be normal had very few abnormal GPRI values or anatomic measurements, whereas the frequency of these abnormalities in IUGR neonates was significantly increased. However, no single anatomic variable was 100% normal in the normal neonates and 100% abnormal in IUGR neonates. Only 40% of IUGR neonates were small for gestational age. Classification of these neonates using principal component analysis and linear discriminant analysis was essentially the same (98.6%) as that made initially after reclassification of two IUGR neonates as normal. The characteristics of the initial and final normal and IUGR groups were very similar and the mNGAS51 was strongly correlated with the NGAS5 in the IUGR group. The effectiveness of mNGASij in separating normal and IUGR neonates increased with the number of GPRIi values included and the types used. GPRIThC and GPRIWT were the most important, followed by GPRIAC. GPRICHL and GPRIHC were much less important and in some cases detrimental. These findings support the concepts of a decrease in soft tissue mass as the initial step in the development of IUGR and the protection of head growth (brain-sparing). The characteristics of mNGAS51, particularly its comprehensiveness, its independence of differences in growth potential, its weighting of GPRIi values according to their importance in the detection of IUGR and its ability to detect different manifestations of IUGR in different individuals, indicate that this should be a most effective parameter for separating normal and IUGR neonates.

Embryonic and Fetal Development↗

PubMed1999

Use of the electrohysterogram signal for characterization of contractions during pregnancy.

This article proposes a method to evaluate the ability of the electrohysterogram signal to characterize the contractions during pregnancy, in a population with high risk of preterm deliveries. This study constitutes a first stage of a project intended to develop a monitoring system for the early diagnosis of preterm deliveries. After a proper signal denoising, we calculate some parameters characteristic of the extracted contractions. These contractions are then divided into classes of different physiological terms. Classical techniques of data analysis, such as principal component analysis and discriminant analysis, permit us to show an evolution of the contractions during pregnancy, which is different between the groups of preterm deliveries and that of deliveries at term. We show that, in an early term of pregnancy, we can separate the two populations: women delivering at term from women delivering preterm. We then show that these two kinds of pregnancy are of different evolutions. These results are encouraging, because they would permit, in a follow-up medical study, to diagnose a possible preterm delivery, as well as the proximity of the delivery.

Adult↗

PubMed1997

A diagnostic index of vertical problems for Class III malocclusions.

A cephalometric investigation was undertaken to derive a vertical indicator--a single value derived from a series of cephalometric variables that can be used in the differential diagnosis of Class III patients. The subjects consisted of 76 Japanese female patients whose chief complaint was anterior crossbite and who presented with either open bite or anterior deep overbite. All the data were derived from the initial lateral cephalometric radiographs. The main skeletal measurements that characterize vertical problems were selected using cluster analysis, principal component analysis, and discriminant analysis. As a result, a Class III vertical indicator was derived. The vertical indicator score for deep overbite was -2.5 +/- 2.2, whereas the vertical indicator score for open bite was +2.5 +/- 2.3. This vertical indicator may be used in diagnosing the nature of vertical problems and determining the appropriate treatment and prognosis for the individual Class III female patient.

Adult↗