Backward Stepwise Discriminant Analysis Research Articles

Applied Multidimensional Analysis for Assessing Youth Performance in Sports Talent Identification Program

This study attempts to apply multidimensional analysis for assessing the profile of male youth in sports talent identification program. Data of anthropometric and physical fitness included power, agility, speeds, flexibility, strength and endurance were obtained from 600 youth in a sports talent identification program aged 13-15 years. Data analyses were carried out using multivariate analysis cluster analysis (CA) and discriminant analysis (DA). Cluster analysis assigned three groups with different profile. While standard mode of DA demonstrated 90.0% accuracy of classification matrix for the assigned groups with nine discriminated variables. Forward and backward stepwise DA discriminated six variables from nine variables with 90.3% level of accuracy. The variables are weight, sitting height, armspan, 20 meter run, 40 meter run and VO2 max. These selected variables of anthropometric and fitness are, therefore, revealed as the essential attributes those must be prioritized for a talent scouting in sports. Present results had demonstrated multidimensional analysis as comprehensive approach capable of providing an information that could help coaches in decision making during youth selection in sports talent identification

Assessment of surface water quality of Pagladia, Beki and Kolong rivers (Assam, India) using multivariate statistical techniques

ABSTRACT Surface water quality monitoring and assessment has become a critical issue because it affects human and aquatic life. In this study, cluster analysis (CA), discriminant analysis (DA) and principal component analysis (PCA) have been applied to identify the possible pollution sources of the Pagladia, Beki and Kolong rivers. Water samples were collected monthly from 27 sampling sites during May 2016 to April 2017. ANOVA analysis showed that there are no statistically significant differences in pollution status in the Pagladia, Kolong and Beki rivers (p > .05). CA was carried out to reveal the similarities among the sampling sites. CA grouped all the sampling sites into two clusters. The first cluster corresponded to the less polluted sampling sites and the second group corresponded to the more polluted sampling sites. Backward stepwise DA showed that T and Ca2+ are the discriminating parameters. PCA was applied on the data set to identify the pollution sources of the surface water. PCA resulted in seven valuable factors for the first cluster, accounting for 90.1% of the total variance, and four valuable factors for the second cluster, accounting for 90.3% of the total variance in water quality data sets. This study illustrated the effectiveness of CA, DA and PCA for a better understanding of large and complex data of surface water quality.

Multi-Hierarchical Pattern Recognition of Athlete’s Relative Performance as A Criterion for Predicting Potential Athletes

: Objective: This study investigates the relative performance quality pattern of athletes that trains under Terengganu sports development program based on physical fitness and psychological components. Methods: Relative performance data (223×7) were obtained from various types of sport, and its main tributaries were evaluated for physical fitness and TEOSQ instrument. Multivariate methods of hierarchical agglomerative cluster analysis (HACA), discriminant analysis (DA), principal component analysis (PCA), and principal factor analysis (PFA), were used to study the relative performance variations of the most significant performance quality variables and to determine the origin of relative performance components. Results: Three clusters of performance were shaped in view of HACA. Forward and backward stepwise DA discriminates six and five performance quality variables from the first seven variables. PCA and FA were used to identify the origin of each quality performance variables based on three clustered groups. Three PCs were obtained with 67% total variation for the highperformance group (HPG) region, three PCs with 72% and 64% total variances were obtained for the moderate-performance group (MPG) and low-performance group (LPG) regions, respectively. The general performance sources for the three groups are from cardiovascular and ego orientation sources. The differences between groups are from flexibility for LPG, task orientation, muscle strength and endurance for MPG and for HPG is flexibility, strength and task orientation. Conclusion: Multivariate methods reveal meaningful information on the relative performance variability of a large and complex athlete’s performance quality data and can be used to determine the significant source and predict potential athletes.Key words: Cluster analysis, Discriminant analysis, Goal orientation, Principal component analysis, Physical fitness, Relative performance.

CHEMOMETRIC ANALYSIS OF SELECTED ORGANIC CONTAMINANTS IN SURFACE WATER OF LANGAT RIVER BASIN

Chemometric techniques namely hierarchical agglomerative cluster analysis (HACA), discriminant analysis (DA), principal component analysis (PCA) and factor analysis (FA) were applied to the distribution of selected organic contaminants (polycyclic aromatic hydrocarbons (PAHs), sterols, pesticides (chlorpyrifos), and phenol) to assess the potential of using these organic contaminants as chemical markers in Langat River Basin. Water samples were collected from February 2012 to January 2013 on a monthly basis for nine monitoring sites along Langat River Basin. HACA was able to classify the sampling sites into three clusters which can be correlated to the level of contamination (low, moderate and high contamination sites). DA was used to discriminate the sources of contamination using the selected organic contaminants and relate to the existing DOE local activities groupings. Forward and backward stepwise DA was able to discriminate two and five organic contaminants variables, respectively, from the original 13 selected variables. The five significant variables identified using backward stepwise DA were fluorene, pyrene, stigmastanol, stigmasterol and phenol. PCA and FA (varimax functionality) were used to identify the possible sources of each organic contaminant based on the inventory of local activities. Five principal components were obtained with 66.5% of the total variation. Result from FA indicated that PAHs (pyrene, fluorene, acenaphthene, benzo[a]anthracene) originated from industrial activity and socio-economic activities; while sterols (coprostanol, stigmastanol and stigmasterol) were associated to domestic sewage and local socio-economic activities. The occurrence of chlorpyrifos was correlated to agricultural activities, urban and domestic discharges. This study showed that the application of chemometrics on the distribution of selected organic contaminants was able to trace the sources of contamination in surface water.

Spatial water quality assessment of Langat River Basin (Malaysia) using environmetric techniques

This study investigates the spatial water quality pattern of seven stations located along the main Langat River. Environmetric methods, namely, the hierarchical agglomerative cluster analysis (HACA), the discriminant analysis (DA), the principal component analysis (PCA), and the factor analysis (FA), were used to study the spatial variations of the most significant water quality variables and to determine the origin of pollution sources. Twenty-three water quality parameters were initially selected and analyzed. Three spatial clusters were formed based on HACA. These clusters are designated as downstream of Langat river, middle stream of Langat river, and upstream of Langat River regions. Forward and backward stepwise DA managed to discriminate six and seven water quality variables, respectively, from the original 23 variables. PCA and FA (varimax functionality) were used to investigate the origin of each water quality variable due to land use activities based on the three clustered regions. Seven principal components (PCs) were obtained with 81% total variation for the high-pollution source (HPS) region, while six PCs with 71% and 79% total variances were obtained for the moderate-pollution source (MPS) and low-pollution source (LPS) regions, respectively. The pollution sources for the HPS and MPS are of anthropogenic sources (industrial, municipal waste, and agricultural runoff). For the LPS region, the domestic and agricultural runoffs are the main sources of pollution. From this study, we can conclude that the application of environmetric methods can reveal meaningful information on the spatial variability of a large and complex river water quality data.

Assessment of temporal and spatial variation of coastal water quality and source identification along Macau peninsula

Cluster analysis (CA), discriminant analysis (DA) and principal component analysis (PCA) were comprehensively coupled to explore and identify the spatial and temporal variation and potential pollution sources in coastal water quality along Macau peninsula. The results show that the 12 months could be grouped into two periods, June–September and the remaining months, and the entire area divided into two clusters, one located at the western sides, and the other on the southeast and southern sides of the Macau peninsula. Through backward stepwise DA, pH, Cl−, TSS, Color and TP, Chloride, Color, NH4 +, DO, COD were discriminant variables of spatial and temporal variation, with 84.82 and 76.54% correct assignments, respectively. Fecal pollution, organic pollution and soil weathering are among the major sources for coastal water quality deterioration along Macau peninsula. This study illustrates that application of multivariate statistical techniques was beneficial to gain knowledge for further optimizing the monitoring network and controlling coastal water quality along Macau peninsula.

Ventilatory Inequalities, Pulmonary Function and Blood Oxygenation in Advanced States of Cystic Fibrosis

In 36 patients with cystic fibrosis (CF), aged 6.2-26.4 years, the relationship between functional lung deterioration and gas exchange characteristics was studied by backward stepwise discriminant analysis of 22 dependent variables, within 3 clinically defined severity degrees of the disease. The stratification was based on a general clinical score, a chest X-ray score and the relative underweight of the patients. Lung function testing included whole body plethysmography and multibreath nitrogen washouts (MBNW). Blood gas analyses were performed under room air (FiO2 = 0.21) and 100% oxygen breathing (FiO2 = 1.0). The clinical score correlated best with vital capacity (r = 0.776) and PaO2 at room air (r = 0.829); the X-ray score correlated best with PaO2 at room air (r = 0.768). Impairment of oxygenation characteristics was closely related to the degree of ventilation inequalities (MBNW) and the amount of trapped gases, both increasingly present from severity group I to III. In addition the calculated arterial-alveolar oxygen ratio (group I 0.47; group II 0.38; group III 0.29) as an estimate of intrapulmonary gas exchange deficiency revealed that only patients with mild lung involvement (group I) fulfil the functional conditions to achieve a paO2 of 70 mm Hg at room air, whereas patients with severer lung involvement (groups II and III) would need a FiO2 of 26.0 or 39.6, respectively, in order to have a sufficient oxygenation.

Self-reported depressive symptoms that differentiate recurrent-episode major depression from dysthymic disorders.

The Beck Depression Inventory (BDI) was administered to 168 outpatients diagnosed with recurrent-episode, major-depression disorders and 99 outpatients diagnosed with dysthymic disorders. The mean ratings of 18 BDI items, except for Sense of Failure, Punishment, and Crying, were comparable in both groups; the mean total-BDI score and mean number of BDI items rated were higher for the major depressive disorders than for the dysthymic disorders. A backward stepwise-discriminant analysis revealed that Suicidal Ideas and Loss of Appetite were the two symptoms that most meaningfully distinguished between the two groups. The major depressive patients described more suicidal ideation and loss of appetite than did the dysthymic patients. The results supported the contention that self-report instruments may be useful in differentiating recurrent-episode major depression from dysthymic disorders.