Search PubMed⌕ Search

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.

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

At least 145 records · Page 8Linked to original sources
PubMed1998

Classification of observational data with artificial neural networks versus discriminant analysis in pharmacoepidemiological studies--can outcome of fluoxetine treatment be predicted?

For several years, there has been an ongoing discussion about appropriate methodological tools to be applied to observational data in pharmacoepidemiological studies. It is now suggested by our research group that artificial neural networks (ANN) might be advantageous in some cases for classification purposes when compared with discriminant analysis. This is due to their inherent capability to detect complex linear and nonlinear functions in multivariate data sets, the possibility of including data on different scales in the same model, as well as their relative resistance to "noisy" input. In this paper, a short introduction is given to the basics of neural networks and possible applications. For demonstration, a comparison between artificial neural networks and discriminant analysis was performed on a multivariate data set, consisting of observational data of 19738 patients treated with fluoxetine. It was tested, which of the two statistical tools outperforms the two other in regard to the therapeutic response prediction from the clinical input data. Essentially, it was found that neither discriminant analysis nor ANN are able to predict the clinical outcome on the basis of the employed clinical variables. Applying ANN, we were able to rule out the possibility of undetected suppressor effects to a greater extent than would have been possible by the exclusive application of discriminant analysis.

Antidepressive Agents, Second-Generation↗
PubMed1986

Discriminant analysis of the clinical indicants for bovine coliform mastitis.

We used discriminant analysis to assess the indicants most useful in predicting whether a cow had coliform bacterial mastitis. One hundred and twenty-nine mastitic cows were divided into two groups, namely those with milk cultures that yielded pure or mixed gram negative organisms, and cows with other organisms or negative culture. Of 21 indicants examined by discriminate analysis only a history of previous mastitis in the affected quarter, weakness, clear or white color of milk, swelling of the udder, water consistency of the milk, lack of previous mastitis in other quarters, lack of palpable udder abscesses, and elevated body temperature were significantly associated with coliform mastitis. Using these variables 78% of cases were correctly classified.

Animals↗
PubMed1985

Structure-activity relationship studies of antiviral N-quinolin-4-yl-N'-benzylidenehydrazine derivatives by discriminant analysis.

The classification technique of linear discriminant analysis (LDA) is applied for studying the structure-activity relationship among antiviral N-quinolin-4-yl-N'-benzylidenehydrazine (II) derivatives. The total hydrophilicity of substituents in the benzylidene moiety along with 4 indicator variables is found to significantly (p less than 0.001) discriminate 25 inactive congeners of (II) from 28 active congeners with more than 80% posterior classification ratio. The predictive stability of the discriminant functions is established by the leave-one-out procedure. In the light of the posterior probabilities of assignment calculated from these functions it is observed that ethoxy group at position 7 and methoxy group at positions 8 and 6 in the quinoline system favour activity while a methoxy group at ortho or para position in the phenyl ring decreases activity. In view of the finer classification within the active class the three-group analysis is also performed using LDA and the adaptive-least-squares techniques.

Antiviral Agents↗
PubMed1988

Discriminant analysis of laboratory tests in patients admitted to a coronary care unit.

Discriminant analysis of chemistry and hematology laboratory test results was used to classify patients with and without myocardial infarction in a coronary care unit. We studied 64 patients with myocardial infarction and 70 patients without infarction, using logistic regression, linear and quadratic discriminant analyses on untransformed and logarithmically transformed data. Serum aspartate aminotransferase (AST, EC 2.6.1.1), the best single discriminating test, classified 73% of patients correctly. Quadratic discriminant analysis on log-transformed data had a 98.5% classification accuracy when all variables were used in the discriminant function and had the highest classification accuracy and precision. All of the discriminant methods had acceptable cross-validation.

Adult↗
PubMed2005

From projection pursuit and CART to adaptive discriminant analysis?

While many efforts have been put into the development of nonlinear approximation theory and its applications to signal and image compression, encoding and denoising, there seems to be very few theoretical developments of adaptive discriminant representations in the area of feature extraction, selection and signal classification. In this paper, we try to advocate the idea that such developments and efforts are worthwhile, based on the theorerical study of a data-driven discriminant analysis method on a simple--yet instructive--example. We consider the problem of classifying a signal drawn from a mixture of two classes, using its projections onto low-dimensional subspaces. Unlike the linear discriminant analysis (LDA) strategy, which selects subspaces that do not depend on the observed signal, we consider an adaptive sequential selection of projections, in the spirit of nonlinear approximation and classification and regression trees (CART): at each step, the subspace is enlarged in a direction that maximizes the mutual information with the unknown class. We derive explicit characterizations of this adaptive discriminant analysis (ADA) strategy in two situations. When the two classes are Gaussian with the same covariance matrix but different means, the adaptive subspaces are actually nonadaptive and can be computed with an algorithm similar to orthonormal matching pursuit. When the classes are centered Gaussians with different covariances, the adaptive subspaces are spanned by eigen-vectors of an operator given by the covariance matrices (just as could be predicted by regular LDA), however we prove that the order of observation of the components along these eigen-vectors actually depends on the observed signal. Numerical experiments on synthetic data illustrate how data-dependent features can be used to outperform LDA on a classification task, and we discuss how our results could be applied in practice.

Algorithms↗
PubMed1992

Factors involved in burn mortality: a multivariate statistical approach based on discriminant analysis.

This article suggests an alternative statistical model for studying the mortality of burned patients: discriminant analysis. This model was applied to our population of 532 patients among whom 71 died. It is not the first time that this model has been applied to assess burn mortality, although it is not frequently used for that application. We found four factors that are statistically significant: age, TBSA, inhalation injury and sex (female). Discriminant analysis allowed us to demonstrate an impressive correlation between death, age and TBSA; inhalation injury by itself and sex, the two other significant factors in our study, seem to have a minor influence on the final outcome of the burned patients and their predictive value is virtually nil. The advantages of this statistical model are compared with logistic regression, the commonly chosen statistical method.

Adult↗
PubMed1991

Generalization of normal discriminant analysis using Fourier series density estimators. Transfusion Safety Study Group.

In this paper we examine the efficiency of a generalization of the traditional normal linear (LDA) or quadratic (QDA) discriminant analysis. This procedure (the generalized discriminant analysis, GDA) replaces each normal density used in the traditional classification rule by a Fourier series density estimator which 'adjusts' the normal density if the data deviate markedly from normality (for example, heavily skewed or multimodal). We derive the GDA in both the univariate and multivariate situations. In a simulation study for the univariate situation, we evaluate the relative efficiency of the GDA. In addition, we demonstrate the performance of the GDA through a series of multivariate applications. We conclude that if the distributions of the data do not deviate markedly from normality, the GDA is as efficient as the LDA or QDA. On the other hand, if either of the distributions deviates from normality, then the GDA, which performs as a semiparametric discriminant procedure, is more efficient than the LDA or QDA.

Antigens, Differentiation↗
PubMed1995

Prediction of the outcome of growth hormone therapy in children with idiopathic short stature. A multivariate discriminant analysis.

OBJECTIVE: To identify, with the use of pretreatment clinical data, the children with idiopathic short stature responsive to treatment with growth hormone (GH). DESIGN: Open, prospective study in a university hospital. SUBJECTS: Patients admitted to the study met the following criteria: birth weight at least 2.5 kg, no sign of dysmorphic disease, stature less than the 3rd percentile for chronologic age (CA), linear growth velocity (GV) less than the 25th percentile for bone age (BA), no sign of puberty, maximal GH response to pharmacologic stimulation greater than 10 micrograms/L, no evidence of organic disease, treatment with daily subcutaneous administration of GH at a dose of 12 to 16 IU/m2 per week. MAIN OUTCOME MEASURES: Eight pretreatment growth variables and the increase of GV after 6 months of therapy were measured. Children with a change in GV that was greater than 2.5 cm/yr after 6 months of GH therapy were considered responders to GH. RESULTS: We studied 67 patients (44 boys). Forty patients (60%) were responders. With univariate analysis the variables found to have predictive value were GV (z score for gender and CA), bone age (z score for gender and CA), and percentage of ideal body weight. These variables were employed in a multivariate discriminant analysis. Growth velocity and BA showed the best independent discriminant analysis. Growth velocity and BA showed the best independent discriminant significance in predicting responsiveness to the initial 6 months of GH therapy. The obtained equation was as follows: Score = -0.40 + 0.92X1 - 0.87X2, where X1 is the GV z value for CA and X2 is the BA z value for CA). Using this scoring system, we obtained a specificity of 96.3% and a sensitivity of 92.5% in predicting responsiveness to GH (chi-square with Yates correction, 48.2; p < 0.001). CONCLUSIONS: Discriminant analysis may permit the pretreatment prediction of responsiveness to the initial 6 months of GH therapy in short children without GH deficiency.

Analysis of Variance↗
PubMed1988

Comparison of discriminant analysis procedures in laboratory differentiation of hypercalcemia.

Logistic, linear, and quadratic discriminant analyses were compared in their ability to differentiate hypercalcemic patients with primary hyperparathyroidism from those with malignancy. Linear and quadratic discriminant analyses were performed by use of both untransformed and logarithmically transformed data. Application of principal components analysis with varimax rotation was helpful in revealing the underlying relationships between variables. All discriminant methods identified serum albumin as the best single discriminating test, with the log-quadratic discriminant analysis classifying 81% of patients correctly. The combination of albumin, carboxy-terminal parathyroid hormone, and chloride improved classification accuracy (92% by use of log-quadratic discriminant analysis). Logistic discriminant analysis, using all 20 variables, gave a classification accuracy of 100%. Quadratic discriminant analysis gave better classification than linear discriminant analysis, and both methods performed better when log-transformed data were used. Logistic discriminant analysis followed by discrimination procedures using log-transformed data yielded the highest classification accuracy and reliability of the methods used.

Humans↗
PubMed1985

[Statistical analysis of a file of angiologic data on a micro- computer. Discriminant analysis of occlusive impedance rheoplethysmography data].

The statistical analysis by multidimensional analysis of the data of a computerized file should lead to the improvement of the quality of the examinations. A discriminant analysis of the data of occlusive impedance rheoplethysmography allows a global analysis of all of the indices, by taking into account both their absolute values and their relative values. This results in a better compromise between the sensitivity and the specificity. The graphic representation gives an immediate conclusion and allows the functional outcome to be monitored.

Arterial Occlusive Diseases↗
PubMed1975

[The differential-diagnostic value of evoked potentials in depressive patients according to discriminant analysis findings].

The authors conducted a discriminant analysis of visual evoked potentials in depressive patients and in normals. The depressive syndrome was correctly diagnosed in 76--90% of the cases. The main role in the diagnosis of such cases belongs to interhemispheric correlational connections in the frontal brain areas. Between these connections there was an intercompensatory relationship. It was possible to demonstrate some traits of intrahemispheric connections in the left hemisphere of normals and in the right hemisphere of the depressed patients. Certain tendencies were marked in the distribution of patients according to the expressiveness of affective disorders.

Adult↗
PubMed1983

Intrahospital prognosis of acute myocardial infarction by means of discriminant analysis: methodological aspects and clinical results.

Discriminant analysis was carried out in 83 patients with acute myocardial infarction who underwent hemodynamic monitoring, to obtain a prognostic index. The classification rate was satisfactory in over 80% of the patients both in a subset used to construct the prognostic index and in a pilot group in which the validity of the index was assessed. In decreasing order of importance the parameters in the function were: age, history of angina, pulmonary artery end-diastolic pressure, diastolic systemic pressure, sum of ST segment elevation, pulmonary vascular resistance, heart rate, history of hypertension and site of necrosis. A much less satisfactory classification rate was obtained using two well-established indexes, perhaps because of sampling, methodological and chronological differences. It thus seems desirable for each hospital to use prognostic indexes obtained from its own population. Such indexes should be updated whenever necessary.

Adult↗
PubMed1986

Selection of effective maximal expiratory parameters to differentiate asthmatic patients from healthy adults by discriminant analysis using all possible selection procedure.

Maximal expiratory volume-time and flow-volume (MEVT and MEFV) curves were drawn for young male nonsmoking healthy adults and for young male nonsmoking asthmatic patients. Eleven parameters, two MEVT (%FVC and FEV1.0%), six MEFV (PFR, V75, V50, V25, V10 and V50/V25), and three MTC parameters (MTC75-50, MTC50-25 and MTC25-RV) were used for the multivariate analysis. The multivariate analysis in this study consisted of correlation coefficient matrix computation, the test for mean values in the multivariates, and the linear discriminant analysis using the all possible selection procedure (APSP). Correlation coefficients among flow rate parameters and flow rate related parameters in high lung volumes were different between the two groups. In the eleven-parameter discriminant analysis by APSP using single parameters, PFR, V75 (flow rate at 75% of forced vital capacity), and FEV1.0% were considered to be the effective parameters. In the seven-parameter discriminant analysis using the parameter groups, the group of all parameters and the %FVC and flow rate-related parameter group were considered to be the effective numerical alternatives to MEFV curves discriminating between healthy adults and asthmatic patients.

Adult↗
PubMed1991

Discriminant analysis of transmission of elevated blood pressure in first generation offspring of African green monkeys.

A linear discriminant analysis was applied to blood pressure data of 162 first generation colony-born offspring of normotensive (C), hypertensive (H), or borderline hypertensive (B) African green monkeys who were being selectively bred in an attempt to establish a strain of spontaneously hypertensive monkeys. The offspring were classified according to their parents' blood pressures as CC, HH, or Mixed (e.g. HC). Blood pressures were measured by indirect methods from unanesthetized offspring aged 0.5-6 years of age. The discriminant score was used to classify each of the 533 blood pressure measurements of the CC, Mixed, and HH offspring into one of three predicted groups: normotensive, borderline hypertensive, or hypertensive. The group means of the three predicted groups compared without regard to offspring type were significantly different (p less than .001). In addition, the percentage of blood pressure measurements predicted to be normal or elevated differed among the three offspring groups (p less than .001). 82% of the blood pressure measurements from CC offspring were classified as normotensive, compared with 58% and 40% of the blood pressure measurements from the Mixed and HH groups, respectively. In contrast, 25% of the blood pressure measurements from the HH groups were classified as hypertensive, compared with 10% and 4% from the Mixed and CC groups, respectively. Blood pressures of the normotensive, borderline hypertensive, and hypertensive subgroups derived from the CC group were consistently and significantly lower (p less than .001) than their respective counterparts in the Mixed and HH groups. The results of the discriminant analysis indicate a trimodal distribution of blood pressures in the first generation offspring and a significant separation of blood pressures among the offspring after a single generation.

Animals↗
PubMed1996

Carcinoembryonic antigen, tissue polypeptide antigen and neuron-specific enolase pleural levels used to classify small-cell and non-small-cell lung cancer patients by discriminant analysis.

The classification of lung cancer into small-cell lung cancer (SCLC) and non-small-cell lung cancer (NSCLC) is essential for disease prognosis and treatment. For this purpose, we have tried to optimize the use of three tumour markers determined on pleural effusions, to differentiate SCLC from NSCLC by means of a canonic variable, generated by discriminant analysis, including subjects with histologically proven lung cancer. Discriminant analysis was performed by using carcinoembryonic antigen, neuron-specific enolase and tissue polypeptide antigen pleural levels, determined in 65 consecutive and unselected patients, histologically classified as 49 NSCLC and 16 SCLC. To validate the formula generated, a control group of 37 lung cancer patients (10 SCLC and 27 NSCLC), enrolled subsequently, was employed. Applying the discriminant analysis to SCLC and NSCLC patients a good classification was obtained (92% rate of correct classification). The aforementioned formula, applied to the validation group, showed a 92% rate of correct classification. This method, which is rapid, inexpensive and routinely applicable to malignant pleural effusions, may be reliably used to classify lung cancer patients.

Adult↗
PubMed2002

An application of linear programming discriminant analysis to classifying and predicting the symptomatic status of HIV/AIDS patients.

This study presents an application of linear programming discriminant analysis (LPDA) to classify and to predict the symptomatic status of HIV/AIDS patients. We applied LPDA as well as several traditional discriminant analysis methods to the AIDS Cost and Services Utilization Survey data set in order to demonstrate the use of LPDA to classify the symptomatic status of HIV/AIDS patients. The potential benefit of LPDA in terms of the classification accuracy was also analyzed.

Diagnosis, Computer-Assisted↗
PubMed2004

Application of linear discriminant analysis to the biochemical and haematological differentiation of opiate addicts from healthy subjects: a case-control study.

OBJECTIVE: Biochemical and haematological parameters of nutritional interest were determined in the serum of opiate addicts in order to compare them with those obtained in healthy subjects. Linear discriminant analysis was applied for the differentiation of the opiate addicts. SUBJECTS: Sera of 106 opiate addicts in detoxification treatment (n=19) or in Methadone Maintenance Treatment Program (MMTP) (n=87) were studied. DESIGN: : The determination of classical biochemical and haematological parameters in blood samples was carried out using standardized methods. Determination of retinol and alpha-tocopherol was carried out by high-performance liquid chromatography with diode-array detector. Folic acid and vitamin B(12) were determined using competitive binding techniques. Minerals were determined by flame emission spectrometry (Na and K) and atomic absorption spectrometry with air-acetylene flame (Ca, Mg, Fe, Cu and Zn) or with hydride generation (Se). Phosphorous was determined using a colorimetric method with ammonium molibdate. All statistical analyses were performed by means of the SPSS version 10.0 software for Windows. RESULTS: Stepwise linear discriminant analysis simplified the system to the following variables: Na, K, Mg, number of leucocytes, triglycerides, GPT, glucose, albumin, retinol and folic acid; and 90.1% (86.4% after crossvalidation) of correct classification was obtained. Representing the first and second discriminant functions, the control groups were well separated from opiate addicts. CONCLUSIONS: Applying linear discriminant analysis on several biochemical and haematological parameters, the opiate addicts could clearly be differentiated from the control individuals, and a tendency to differentiate the opiate addicts in MMTP and in detoxification treatment was observed.

Adult↗
PubMed2005

Discriminant analysis of the cognitive performance profile of MS patients differentiates their clinical course.

OBJECTIVE: To compare the neuropsychological deficits of primary progressive multiple sclerosis with those of relapsing-remitting and secondary progressive multiple sclerosis. METHODS: Sixty-five patients with different clinical courses of MS were neuropsychologically tested for language, attention, memory and executive functions. Discriminant analysis was used to predict the type of clinical course either by clinical variables (age, EDSS and duration of illness) or neuropsychological test results. RESULTS: For single neuropsychological tests, group differences were rare between the progressive courses and the relapsing-remitting course of MS or absent between the progressive courses of MS. However, discriminant analysis correctly identified 87.7 percent of the patients' courses in general, and about 90 percent of the patients with chronic progressive MS. CONCLUSION: Using discriminant analysis, this study found neuropsychological impairment characteristic for relapsing remitting, secondary progressive and primary progressive patients.

Adult↗