Canonical Discriminant Function Coefficients Research Articles

Age dynamics of functional coordination and force readiness girls grades 8-9

Purpose — to identify patterns of age dynamics of functional and motor fitness girls eighth and ninth grades.Material and methods. To achieve the objectives adopted the following methods: analysis of scientific literature, teacher testing and treatment methods of mathematical statistics research results. In the study involved 14 dyvchat class 8, 14 — 9 class.Conclusions. Analysis of the test results shows that statistically significant differences between girls 8 and 9 classes in functional samples was observed. For the functional state of respiration and circulation girls 8 and 9 classes are rated as healthy untrained.Girls 9 classes show statistically significantly better results in tests for speed strength, static display of strength and motor coordination (tests 8 «Long jump from place, see» 7 «Height in folded hands, p ‘and 2’ Evaluation of temporal parameters movement, with error «).Standardized canonical discriminant function coefficients can determine the value contribution of variables in the function result. With the largest contribution to the canonical function variables are 6, 10 and 7: the higher the value of these variables, the more important functions. The foregoing indicates to assess the functional and motor fitness of students grades 8—9 using the proposed battery of tests.Structural factors canonical discriminant function shows that feature most significantly associated with 7, 8 and 2 variables: the more attention is paid to static, the actual strength and coordination training, the greater the likelihood of increasing the level of functional and motor fitness girls 8—9 grades.100.0% initial observations are grouped correctly classified. Thus, the canonical discriminant function can be used for evaluation and prediction of functional and motor fitness girls 8—9 grades.

Determination of the discriminant score of intestinal microbiota as a biomarker of disease activity in patients with ulcerative colitis

BackgroundIn recent years, the gut microbiota has been found to provide an important link to the development of inflammatory bowel diseases (IBD) like ulcerative colitis (UC). Accordingly, inter-individual variation in the gut microbial community may be linked to inter-individual variation in the risk of IBD or other diseases. Further, the Terminal Restriction Fragment Length Polymorphism (T-RFLP) is a molecular biology technique for profiling bacterial species in faecal samples. This study was to evaluate a biomarker based on intestinal microbiota.MethodsThe study subjects were 69 patients with UC together with 80 relatives as controls. Twenty-three patients had active UC (group I) and 46 had quiescent UC (group II). The later included 17 patients with mild inflammation in the large intestine (group IIa), 29 without inflammation (group IIb). The patients’ relatives were consanguineous (group III, n = 47), and non-consanguineous (group IV, n = 33). Faecal samples were obtained from all subjects for the investigation of intestinal microbiota by applying the T-RFLP method. The Discriminant analysis of operational-taxonomic-unit (OTU) on T-RFLP fingerprints was performed. The Canonical Discriminant Function Coefficient (Df) for each OTU was calculated. The individual OTUs were multiplied by the Df value, and the sum was termed the Discriminant Score (Ds).ResultsThe Ds decreased thus: group I > group IIa > group IIb > group III > group IV. Significant difference was calculated for group I vs group IV (P < 0.01), group I vs group IIb (P < 0.05), group I vs group III (P < 0.01), group IIa vs IV (P < 0.01), group IIb vs group IV (P < 0.01), group III vs group IV (P < 0.01), indicating a strong association between gut microbial species and the development of UC.ConclusionsIn this study, the Ds related to UC, or otherwise absence of UC in the five groups. Potentially, Ds may become a clinically relevant biomarker of disease activity in UC. To our knowledge, this is the first application of the Ds to the study of microbiota in UC patients, consanguineous and non-consanguineous relatives.Trial registrationClinical trial No: UMIN 000004123.

Su1185 New Test for Profiling of the Gut Microbiota in Non-IBD and IBD Patients

to a labeled end of an amplified gene. T-RFLP is an established technique for analysing bacterial species in faecal samples. Methods: This was a quantitative analysis with a major focus on evaluating a biomarker based on intestinal microbiota. The subjects were 69 UC patients together with an 80 relatives as controls. Twenty-three patients had active UC (group I) and 46 had quiescent UC (group II). The latter included 17 with mild inflammation in the large intestine (group IIa), and 29 without inflammation (group IIb). The patients’ relatives were consanguineous (group III, n = 47), and non-consanguineous (group IV, n = 33). The Discriminant analysis of operationaltaxonomic-unit (OTU) on T-RFLP fingerprints was done. The Canonical Discriminant Function Coefficient (Df) for each OTU was calculated by using group I and group IV, in whom the difference of intestinal flora was thought to be greatest. The individual OTUs were multiplied by the Df value, and the sum was termed the Discriminant Score (Ds). The analysis was by SPSS (IBM Statistics 20.0). Results: The Ds decreased thus: group I > group IIa > group IIb > group III > group IV. As certain strains of bacteria are associated with the development of UC, while others not, this should show up in the OTU values in T-RFLP analysis. The analysis was set to factor this assumption, and the OUTs with 0 value 95% of all cases together with the OTUs of structure matrix value <0.01 according to the Discriminant analysis were excluded. Significant difference was calculated for group I vs. group IV (P< 0.01), group I vs. group IIb (P< 0.05), group I vs. group III (P< 0.01), group IIa vs. group IV (P< 0.01), group IIb vs. group IV (P< 0.01), group III vs. group IV (P< 0.01). Conclusions: In this study, the Ds related to UC activity, or otherwise absence of UC in the five groups. Potentially, Ds may become a clinically relevant biomarker of disease activity in UC. To our knowledge, this is the first application of the Ds to the study of microbiota in UC patients, consanguineous and non-consanguineous relatives.

The role of shame in distinguishing perpetrators of intimate partner violence in U.S. veterans

Increasing attention is being paid to the fact that exposure to traumatic stressors in military combat may lead to perpetration of intimate partner violence (IPV). Because shame has been identified as a factor in posttraumatic stress disorder (PTSD), the current cross-sectional study examined the relationship in U.S. veterans between IPV and PTSD, depression, guilt, and shame. We hypothesized that shame would be the strongest correlate of perpetration of IPV and that shame would mediate the relationship between PTSD and IPV. Participants were 264 primarily male and Caucasian mixed-era veterans presenting for psychological treatment at a Veterans Affairs hospital. They completed standard measures of depression, PTSD symptoms, shame, and guilt and a local checklist was used to dichotomize the sample regarding IPV. Discriminant analysis indicated that shame contributed most (standardized canonical discriminant function coefficient = .44) to distinguishing perpetrators of IPV. In addition, the results were consistent with shame as a mediator of the relationship between PTSD and IPV. These results are in line with studies indicating that shame is linked to IPV perpetration in nonveteran samples (Harmon, 2002; Rand, 2004; Schibik, 2002) and suggests that shame may be an important aspect of the relationship between PTSD and IPV.

Infant and Under-five Mortality in Bangladesh: Discriminant Analysis

Abstract Bangladesh is on target for achieving the Millennium Development Goal 4 relating to infant and under-five mortality because of very rapid reduction in mortality in recent years. But this rate of reduction may be difficult to sustain and may hamper the achievement of Millennium Development Goal 4. Therefore, the main objective of this paper is to discuss and compare the different covariates of infant and under-five mortality in the context of overall country, urban and rural levels of Bangladesh using discriminant analysis. For this, the data are taken from Bangladesh Demographic and Health Survey, 2004. In discriminant analysis, the stepwise procedure has been picked up and only the significant variables are ranked according to the rank of Wilk’s Lambda values. The canonical discriminant function coefficients (unstandardized and standardized) for the predictor variables have also been calculated. Both the results show that breastfeeding is the most important variable in discriminating the two groups of mothers, i.e., mothers experiencing to infant mortality or not and mothers experiencing to under-five mortality or not. The related results of discriminant function also indicate that the discriminant function is statistically significant and discriminates well. Therefore, improvements in the health system are essential for promoting the breastfeeding practices (both inclusive and exclusive), which may be the effective strategies to reach families and communities with targeted messages and information.

Time and frequency parameters of bottlenose dolphin whistles as predictors of surface behavior in the Mississippi Sound

Acoustic characteristics related to contour of the whistle (such as highest and lowest frequency, beginning and ending frequency, whistle duration, and number of turns) of bottlenose dolphin (Tursiops truncatus) whistles were measured to test whether any of the measurements were related to the behavioral state of the dolphins when the whistle was recorded (coded as mill, travel, mill/travel, feed, or social). Objective measures of time and frequency were obtained using Raven, while number of turns in a whistle was determined by human raters. In all a series of discriminant function analyses using the acoustic characteristics to predict the behavioral state, the highest standardized canonical discriminant function coefficients were: lowest frequency, number of turns, and duration. The models that incorporated these variables performed significantly better than chance at correctly assigning the whistles into the surface behavior category in which they were recorded. The rate of whistling was related to group size, surface behavior and season via a series of two-way ANOVAs (analysis of variance).

Accuracy of 3-Tesla magnetic resonance imaging for the staging of prostate cancer in comparison to the partin tables

An accurate prediction of final pathological stage is essential for proper patient selection to determine who will experience maximum benefit from radical prostatectomy. In this context, the Partin tables represent one of the most widely used statistical prediction tools. To compare the accuracy in predicting pathological stage in patients intended for radical prostatectomy between 3-Tesla (T) magnetic resonance imaging (MRI) and the Partin tables in a prospective trial. 27 men with staging results from 3T MRI using a phased-array coil were compared in terms of staging accuracy with whole-mount-section histopathologic analyses as the standard of reference. Probabilities for pathological stages were estimated according to the Partin tables. Spearman rank correlation and discriminant analysis were calculated to assess relationships. Histopathological evaluation revealed organ-confined disease (OC) in 21 (77.8%) and extracapsular extension (ECE) in six (22.2%) men. Three-Tesla MRI staging was accurate in all patients with OC and in four out of the six men with ECE. Accuracy of local staging was 85.2% (23 of 27). Sensitivity was 66.7% (95% confidence interval [CI] 0.223-0.957) and specificity 100% (95% CI 0.839-1) for the detection of ECE. Findings of MRI and the Partin tables showed a Spearman rho of 0.780 vs. 0.254 for OC and 0.780 vs. 0.363 for ECE, respectively. Compared to the Partin tables, MRI revealed standardized canonical discriminant function coefficients of 0.992 (P<0.001) vs. 0.205 (P=0.432) for OC and 0.965 (P<0.001) vs. 0.329 (P=0.197) for ECE, respectively. 3T MRI showed a high accuracy for the staging of clinically localized prostate cancer, and it was significantly more accurate in predicting the final pathological stage than the Partin tables.

How to Increase the Diagnostic Value of Malignancy-Related Ascites: Discriminative Ability of the Ascitic Tumour Markers

Making a differential diagnosis between malignant and non-malignant ascites is an important clinical issue, but cytological examination has a relatively low diagnostic sensitivity. This study aimed to find a discriminative model that distinguished between malignancy-related and non-malignant ascites. The study included 107 patients: 50 with non-malignant and 57 with malignant ascites. Ascites was analysed using a range of tumour markers and standard cytology. Standardized canonical discriminant function coefficients were used to distinguish between ascites types. The combination of carbohydrate antigen (CA) 15-3, carcinoembryonic antigen (CEA) and cytokeratin 19 fragments (CYFRA-21.1) discriminated between malignancy-related ascites and non-malignant ascites with an accuracy of 98.8% compared with an accuracy of 77.8% for cytological examination. In conclusion, the use of a discriminant function constructed from a combination of CA15-3, CEA and CYFRA-21.1 could distinguish malignant from non-malignant ascites with greater accuracy than cytological examination. Further studies in larger population groups are warranted.

판별분석을 이용한 한강권역 농업용 하천수의 수질등급모형

본 연구는 한강권역 농업용 소하천 88지점에 대하여 현재의 수질상태를 고려하여 수질등급모형을 구축하고자 실시하였다. 10개의 수질항목에 대하여 각 수질항목별 상위 20%씩 level을 부여한 후, 최저 등급을 기준으로 전체 수질등급을 부여하였는데, Level I에 해당하는 하천은 없었고, Level II는 1지점, Level III은 3지점, Level IV는 22지점, Level V는 62지점으로 조사되어 등급이 낮은 Level V로 갈수록 많은 하천들이 속하였다. 집단간 차이의 유무를 알아보기 위해 집단 평균의 동질성을 검정한 결과, DO, BOD, <TEX>$COD_{Cr},\;NH_3-N$</TEX>, 그리고 SS가 집단 간에 차이가 없는 것으로 조사되어 판별분석 시에는 제외되었다. 판별력이 가장 뛰어난 <TEX>$NO_3-N$</TEX>가 수질등급에 관여하는 변수로 선택되었으며, 분류함수계수에 의한 각 수질등급별 등급모형은 Level II=<TEX>$-4.648+3.246{\times}[NO_3-N]$</TEX>, Level III=<TEX>$-5.084+3.456{\times}[NO_3-N]$</TEX>, Level IV=<TEX>$-4.298+3.067{\times}[NO_3-N]$</TEX>, Level V=<TEX>$-7.369+4.396{\times}[NO_3-N]$</TEX>로 구축되었고, 적합도 평가에 의한 구축된 수질 등급모형의 적합도는 88.4%로 조사되어 높은 적합도를 보였지만, 향후 더 많은 자료의 확보와 축적으로 수질등급모형이 수정, 보완된다면 활용가능성은 더 높을 것으로 판단되며, 이를 바탕으로 각 권역별로도 실제 수질에 적합한 수질등급모형을 구축할 수 있으리라 생각된다. The main purpose of this work is the development of water quality level model using the data such as DO, EC, BOD, <TEX>$COD_{Cr},\;NH_3-N,\;NO_3-N,\;PO_4-P$</TEX>, T-N, T-P, and SS in 88 agricultural streams of the Han river watersheds. To grant water quality level for each parameters, it divided into 20% respectively in the order of water quality level. On the basis of the lowest water quality level, water quality of streams was assigned. As the result, number of stream corresponding to Level Ⅰ was 0, Level II was 1 stream, Level III was 3 streams, Level IV was 22 streams, and Level V was 62 streams. By standardized canonical discriminant function coefficient, <TEX>$NO_3-N$</TEX> was the highest in 0.427 at the discriminant power. According to discriminant function for water quality level, it was equal to <TEX>$-4.648+3.246{\times}[NO_3-N],\;-5.084+3.456{\times}[NO_3-N],\;-4.298+3.067{\times}[NO_3-N],\;and\;-7.369+4.396{\times}[NO_3-N]$</TEX> from Level II to Level V, respectively. As a result of test at real data of the Han river watersheds in 2007, the suitability of water quality level model was high to 88.4%.

Verification test of sensory analyses of comb and strained honeys produced as pure and feeding intensively with sucrose (Saccharum officinarum L.) syrup

The aims of the study were to discriminate comb and strained honeys produced by the standard beekeeping method (control), shaking method (pure blossom honey), and feeding intensively (100kg/colony) with sucrose (adulterated honey) syrup by using sensory analysis and to develop a method to be used in identification of unknown or suspicion honey samples. In the study, twenty trained panelists assessed honey samples in relation to their properties including taste, odor, color, aroma, viscosity, dissolution in mouth, inflammation in throat, attractiveness, flavor and general impression during four months. There were no differences in odor, viscosity, and dissolution in mouth between comb and strained honey samples which produced by different methods (P > 0.05). Discrimination of strained honey by sensory analysis was more reliable when compared to comb honey. The ratio of correctly classified sample was 78.3% for comb and 86.7% for strained honey. The more honey was pure the more discrimination of honey sample by sensory analysis was reliable. In verification test five unknown honey samples were classified 100% in their real groups by using canonical discriminant function Coefficients of each properties evaluated and the projections of the sample points on the plane of the canonical function-1 and function-2.

Wisconsin Card Sorting Test Performance after Putaminal Hemorrhagic Stroke

The purpose of this study was to compare the performance differences in the Wisconsin Card Sorting Test (WCST) between 55 patients with putaminal hemorrhage (PH) 3 months after stroke and 69 age-matched normal controls. Impairment on WCST was defined as performance greater than 1.64 standard deviation below the control mean. A multivariate analysis of covariance (MANCOVA) controlling for education yielded a significant main effect for group but not for education and interaction of group x education. Univariate analyses revealed significant between-group differences in five WCST measures, including perseverative errors (PE), perseverative responses (PR), conceptual-level responses (CLR), number of categories completed (NCC), and trials to complete the first category (TCC). For patients with PH, z-scores for two WCST indices were within the impaired range: TCC and PR. A high percentage of patients (40-47%) scored in the designated impaired range on NCC, PR, PE, and TCC. The WCST variables discriminated patients from controls with an overall accurate classification rate of 91.9%. Of these, the variables that contributed most to the differentiation between patients and normal controls were PE, CLR, and total number correct (TNC) (a standardized canonical discriminant function coefficient > 0.40). Finally, no significant hemispheric laterality effects emerged on any of the WCST variables. The results of this study provide further evidence of impaired mental set shifting in stroke patients with PH. The implications for rehabilitation professionals are discussed, and recommendations for further research are made.

Heterogeneity of DNA distribution in diploid cells: a new predictive discriminant factor for solid tumour behaviour.

Spatial nuclear DNA heterogeneity distribution of Feulgen‐stained DNA diploid cells was studied by image cytometry in voided urine of 119 patients without bladder tumour (n=20) and with initial (n=23) or previous (n=76) diagnosed bladder tumour. For each patient, repetitive DNA measurements were performed during 1–4 years of follow up. Only cells of diploid DNA histograms and diploid subpopulations of aneuploid DNA histograms were used for analysis. DNA heterogeneity distribution of these diploid cells was quantified by statistical parameters of each nuclear optical density distribution. Discriminant analysis was performed on three groups of DNA histograms. Group A (n=44): aneuploid DNA histograms of patients with bladder tumour. Group D (n=55): 38 diploid DNA histograms of the 20 patients without bladder tumour (subgroup D1) and 17 diploid DNA histograms of patients with a non‐recurrent bladder tumour (subgroup D2). Group R (n=27): diploid DNA histograms of patients with bladder tumour recurrence. No statistically significant discriminant function was found to separate D1 and D2. However, the first canonical discriminant function C1 differentiated diploid cells of diploid DNA histograms (group D and group R) from diploid cell subpopulations of aneuploid DNA histograms (group A). Mean C1 values were 1.06, 0.84 and –1.45 for groups R, D and A, respectively. The second canonical discriminant function C2 differentiated diploid DNA histograms of patients with bladder tumour recurrence (group R) from diploid DNA histograms of patients without bladder tumour or without bladder tumour recurrence (group D). Mean C2 values were 1.78 and –0.76 for groups R and D, respectively. In 95% confidence limit, the rate of rediscrimination using the two first canonical discriminant functions C1 and C2 were 86.4, 74.5 and 74.1% for groups A, D and R, respectively. Percent of “grouped” cases correctly classified was 78.6%. Thus spatial DNA heterogeneity distribution of diploid cells seems to quantitate probable genetic instability as a function of clinical evolution such as tumour recurrence, and suggests the possible presence of aneuploid stemlines in a heterogeneous tumour, even if a diploid DNA histogram is observed in a single sample. From standardized C1 and C2 canonical discriminant function coefficients, a DNA heterogeneity index (2c‐HI) is proposed to characterize diploid cells providing a descriptive and predictive discriminant factor for solid tumour behaviour.

Standardized Discriminant Coefficients: A Rejoinder

The purpose of this article is to respond to a recent paper by Nordlund and Nagel (1991) who took issue with recommendations made by the present authors regarding the standardization of canonical discriminant coefficients (Mueller & Cozad, 1988). We welcomed their insightful comments and clarifications but are not convinced that their arguments are sufficient to accept the recommendation of using total variance estimates to standardize canonical discriminant function coefficients.

Assessing the contribution of individual variables following rejection of a multivariate hypothesis

There are two basic approaches for examining the contribution of in-dividual variables to separation of groups after rejection of a multivariate hypothesis: (1) a multivariate approach showing the contribution of each variable in the presence of the other variables, and (2) a univariate approach showing the contribution of each variable by itself ignoring the other vari-ables. For the multivariate approach, we express the standardized (canonical) discriminant function coefficients in a form showing that the contribution of each variable is due to its multiple correlation with the other variables and how well its separation of groups can be predicted from the other variables. For the univariate approach, we present the results of a Monte Carlo study comparing four methods of testing individual variables. A “protected” pro-cedure that performs F tests only if the overall Wilks' A rejects, appears to be preferred.

Relationship between diet-related behavior and caries in a group of 14-year-old Danish children.

Dietary factors related to dental caries may be both biological and behavioral. In the present study the relative contribution of a number of behavioral characteristics is analyzed in order to differentiate between caries-active (CA) and caries-inactive (CI) children. The study group comprised 49 CA and 55 CI 14-yr-old Danish children, constituting the dichotomous dependent variable. Several variables related to the social and cultural context in which food and snack consumption takes place were studied. Subsequently, their relative contributions to the discriminant function (CA vs. CI) were estimated using discriminant analysis. The three strongest discriminating variables were "weekly milk consumption" (-0.726), "packed lunch from home" (0.554), and "weekly consumption of candy in DKK" (0.306) as measured by standardized canonical discriminant function coefficients. These variables may be indicative of the impact of the social environment on dietary habits, pointing to the need for incorporating the social network as reinforcing agent in dietary counseling if dietary behavior is to be changed.