Home About us Contact
|
Home About us Contact | ||
|
|
||
Backward Elimination (backward + elimination)
Selected AbstractsPlant species richness of nature reserves: the interplay of area, climate and habitat in a central European landscapeGLOBAL ECOLOGY, Issue 4 2002Petr Py Abstract Aim To detect regional patterns of plant species richness in temperate nature reserves and determine the unbiased effects of environmental variables by mutual correlation with operating factors. Location The Czech Republic. Methods Plant species richness in 302 nature reserves was studied by using 14 explanatory variables reflecting the reserve area, altitude, climate, habitat diversity and prevailing vegetation type. Backward elimination of explanatory variables was used to analyse the data, taking into account their interactive nature, until the model contained only significant terms. Results A minimal adequate model with reserve area, mean altitude, prevailing vegetation type and habitat diversity (expressed as the number of major habitat types in the reserve) accounted for 53.9% of the variance in species number. After removing the area effect, habitat diversity explained 15.6% of variance, while prevailing vegetation type explained 29.6%. After removing the effect of both area and vegetation type, the resulting model explained 10.3% of the variance, indicating that species richness further increased with habitat diversity, and most obviously towards warm districts. After removing the effects of area, habitat diversity and climatic district, the model still explained 9.4% of the variance, and showed that species richness (i) significantly decreased with increasing mean altitude and annual precipitation, and with decreasing January temperature in the region of the mountain flora, and (ii) increased with altitudinal range in regions of temperate and thermophilous flora. Main conclusions We described, in quantitative terms, the effects of the main factors that might be considered to be determining plant species richness in temperate nature reserves, and evaluated their relative importance. The direct habitat effect on species richness was roughly equal to the direct area effect, but the total direct and indirect effects of area slightly exceeded that of habitat. It was shown that the overall effect of composite variables such as altitude or climatic district can be separated into particular climatic variables, which influence the richness of flora in a context-specific manner. The statistical explanation of richness variation at the level of families yielded similar results to that for species, indicating that the system of nature conservation provides similar degrees of protection at different taxonomic levels. [source] Using feedforward neural networks and forward selection of input variables for an ergonomics data classification problemHUMAN FACTORS AND ERGONOMICS IN MANUFACTURING & SERVICE INDUSTRIES, Issue 1 2004Chuen-Lung Chen A method was developed to accurately predict the risk of injuries in industrial jobs based on datasets not meeting the assumptions of parametric statistical tools, or being incomplete. Previous research used a backward-elimination process for feedforward neural network (FNN) input variable selection. Simulated annealing (SA) was used as a local search method in conjunction with a conjugate-gradient algorithm to develop an FNN. This article presents an incremental step in the use of FNNs for ergonomics analyses, specifically the use of forward selection of input variables. Advantages to this approach include enhancing the effectiveness of the use of neural networks when observations are missing from ergonomics datasets, and preventing overspecification or overfitting of an FNN to training data. Classification performance across two methods involving the use of SA combined with either forward selection or backward elimination of input variables was comparable for complete datasets, and the forward-selection approach produced results superior to previously used methods of FNN development, including the error back-propagation algorithm, when dealing with incomplete data. © 2004 Wiley Periodicals, Inc. Hum Factors Man 14: 31,49, 2004. [source] Spectral analysis of parallel incomplete factorizations with implicit pseudo-overlapNUMERICAL LINEAR ALGEBRA WITH APPLICATIONS, Issue 1 2002M. Magolu monga Made Abstract Two general parallel incomplete factorization strategies are investigated. The techniques may be interpreted as generalized domain decomposition methods. In contrast to classical domain decomposition methods, adjacent subdomains exchange data during the construction of the incomplete factorization matrix, as well as during each local forward elimination and each local backward elimination involved in the application of the preconditioner. Local renumberings of nodes are combined with suitable global fill-in strategy in an (successful) attempt to overcome the well-known trade-off between high parallelism (locality) and fast convergence (globality). From an algebraic viewpoint, our techniques may be implemented as global renumbering strategies. Theoretical spectral analysis is provided, which displays that the convergence rate weakly depends on the number of subdomains. Numerical results obtained on a 16-processor SGI Origin 2000 are reported, showing the efficiency of our parallel preconditionings. Copyright © 2001 John Wiley & Sons, Ltd. [source] Validation of whole genome linkage-linkage disequilibrium and association results, and identification of markers to predict genetic merit for twinningANIMAL GENETICS, Issue 4 2010C. D. Bierman Summary A previous genome-wide search with a moderate density 10K marker set identified many marker associations with twinning rate, either through single-marker analysis or combined linkage-linkage disequilibrium (LLD; haplotype) analysis. The objective of the current study was to validate putative marker associations using an independent set of phenotypic data. Holstein bulls (n = 921) from 100 paternal half-sib families were genotyped. Twinning rate predicted transmitting abilities were calculated using calving records from 1994 to 1998 (Data I) and 1999 to 2006 (Data II), and the underlying liability scores from threshold model analysis were used as the trait in marker association analyses. The previous analysis used 201 bulls with daughter records in Data I. In the current analysis, this was increased to 434, providing a revised estimate of effect and significance. Bulls with daughter records in Data II totaled 851, and analysis of this data provided the validation of results from analysis of Data I. Single nucleotide polymorphisms (SNPs) were selected to validate previously significant single-marker associations and LLD results. Bulls were genotyped for a total of 306 markers. Nine of 13 LLD regions located on chromosomes 1, 2, 3, 6, 9, 22, 23(2) and 26 were validated, showing significant results for both Data I and II. Association analysis revealed 55 of 174 markers validated, equating to a single-marker validation rate of 31%. Stepwise backward elimination and cross-validation analyses identified 18 SNPs for use in a final reduced marker panel explaining 34% of the genetic variation, and to allow prediction of genetic merit for twinning rate. [source] Exploring the role of face processing in facial emotion recognition in schizophreniaACTA NEUROPSYCHIATRICA, Issue 6 2009Paola Rocca Objective: Impairment in emotion perception represents a fundamental feature of schizophrenia with important consequences in social functioning. A fundamental unresolved issue is the relationship between emotion perception and face perception. The aim of the present study was to examine whether facial identity recognition (Identity Discrimination) is a factor predicting facial emotion recognition in the context of the other factors, known as contributing to emotion perception, such as cognitive functions and symptoms. Methods: We enrolled 58 stable schizophrenic out-patients and 47 healthy subjects. Facial identity recognition and emotion perception were assessed with the Comprehensive Affect Testing System. Different multiple regression models with backward elimination were performed in order to discover the relation of each significant variable with emotion perception. Results: In a regression including the six significant variables (age, positive symptomatology, Identity Discrimination, attentive functions, verbal memory-learning, executive functions) versus emotion processing, only attentive functions (standardised , = 0.264, p = 0.038) and Identity Discrimination (standardised , = 0.279, p = 0.029) reached a significant level. Two partial regressions were performed including five variables, one excluding attentive functions and the other excluding Identity Discrimination. When we excluded attentive functions, the only significant variable was Identity Discrimination (standardised , = 0.278, p = 0.032). When we excluded Identity Discrimination, both verbal memory-learning (standardised , = 0.261, p = 0.042) and executive functions (standardised , = 0.253, p = 0.048) were significant. Conclusions: Our results emphasised the role of face perception and attentional abilities on affect perception in schizophrenia. We additionally found a role of verbal memory-learning and executive functions on emotion perception. The relationship between those above-mentioned variables and emotion processing could have implications for cognitive rehabilitation. [source] | ||||||||||