Home About us Contact
|
Home About us Contact | ||
|
|
||
Weight Regressions (weight + regression)
Selected AbstractsGeographically Weighted Discriminant AnalysisGEOGRAPHICAL ANALYSIS, Issue 4 2007Chris Brunsdon In this article, we propose a novel analysis technique for geographical data, Geographically Weighted Discriminant Analysis. This approach adapts the method of Geographically Weighted Regression (GWR), allowing the modeling and prediction of categorical response variables. As with GWR, the relationship between predictor and response variables may alter over space, and calibration is achieved using a moving kernel window approach. The methodology is outlined and is illustrated with an example analysis of voting patterns in the 2005 UK general election. The example shows that similar social conditions can lead to different voting outcomes in different parts of England and Wales. Also discussed are techniques for visualizing the results of the analysis and methods for choosing the extent of the moving kernel window. [source] Modélisation spatiale de la pauvretéà Montréal: apport méthodologique de la régression géographiquement pondéréeTHE CANADIAN GEOGRAPHER/LE GEOGRAPHE CANADIEN, Issue 4 2007PHILIPPE APPARICIO Spatial Modeling of Poverty in Montréal: Methodological Contribution of the Geographically Weighted Regression The Island of Montréal is particularly concerned with the issue of poverty. In 2000, 29 percent of its inhabitants lived under the low income cut-offs as defined by Statistics Canada. However, poverty is not a homogeneous phenomenon at the intra-urban scale, and identifying and categorizing spaces of poverty has become a main concern for ongoing researches. According to this way of thinking, this paper proposes an analysis of the factors influencing the geographical distribution of poverty on the Island of Montréal. To be able to identify properly the various profiles of poverty, this analysis uses a specific methodology, the geographically weighted regression (GWR), and compares its results with the ones of a classical regression model. At the global level, the most important factors to explain poverty are in order: unemployment, lone-parent families, one person households, recent immigrants, part time or part year workers, school dropouts. At the local level, L'île de Montréal est particulièrement touchée par la pauvreté, puisqu'en 2000 29 pour cent de sa population vivait sous le seuil de faible revenu tel que défini par Statistique Canada. La pauvreté ne constituant pas toutefois un phénomène homogène à l'échelle intra-urbaine, l'identification et la qualification des zones de pauvreté deviennent des enjeux de recherche de première importance. Dans cette perspective, cet article propose une analyse des facteurs qui déterminent la distribution spatiale de la pauvreté au niveau des secteurs de recensement de l'île de Montréal. Pour ce faire, l'analyse mobilise un outil méthodologique particulier: la régression géographiquement pondérée, et en compare les résultats avec un modèle de régression multiple global. Au niveau global, on constate que les facteurs classiques conduisant à la pauvreté sont à l',uvre sur le territoire de l'île de Montréal. Dans l'ordre, ces facteurs sont: le chômage, la monoparentalité, le fait de vivre seul, le fait d'être un immigrant récent, le travail atypique et la non-fréquentation scolaire des jeunes de 15 à 24 ans. Au niveau local, s'il est vrai we observe that variables employment and lone-parents families play significantly in almost all the census tracts, the four other factors are significant only in some census tracts in the center of the Island. At the end of this analysis, the advantages of the GWR methodology appear clearly, as its capacity to take into account the geographical variations of the phenomenon allows a better identification and categorization of poverty areas in Montréal. que le chômage et la monoparentalité agissent significativement dans presque tous les secteurs, les quatre autres facteurs sont uniquement significatifs dans certains secteurs du centre de l'île. Au terme de l'analyse, les avantages de la régression géographiquement pondérée apparaissent clairement, sa plus grande sensibilité aux variations spatiales du phénomène permettant de mieux identifier et qualifier les zones de pauvreté montréalaises. [source] The influence of different single dietary sources on moult induction in laying hensJOURNAL OF THE SCIENCE OF FOOD AND AGRICULTURE, Issue 14 2007Behzad Mansoori Abstract An investigation was carried out to assess the possibility of using single dietary sources as alternatives to feed deprivation for the induction of moult in commercial laying hens. The study involved six dietary groups of 29 laying hens: unmoulted, dried tomato pomace, alfalfa meal, rice bran, cumin seed meal and feed withdrawal. The birds received the above diets during the moulting period (11 days), and body weight loss and ovary weight regression were measured. Post-moult production parameters (number of eggs produced per hen per day, egg weight, shell weight, yolk colour and Haugh unit) were measured for 12 weeks. Results showed that all dietary sources were as effective as feed withdrawal in causing ovary weight regression in birds. Birds provided with tomato pomace or alfalfa showed lower weight losses than feed-deprived birds at the end of the moulting period. Hens moulted by tomato pomace or alfalfa exhibited post-moult levels of egg production over a 12 week period that were superior to those of hens moulted by feed withdrawal. Post-moult eggs laid by hens moulted by all dietary sources were of comparable quality to eggs from feed-deprived hens and superior to those from unmoulted hens. As fibrous feeds with low metabolisable energy and an appreciable amount of protein, dried tomato pomace and alfalfa meal may be fed to hens on an ad libitum basis for effective moult induction while reducing the stress of severe starvation and retaining comparable egg quality and production. Copyright © 2007 Society of Chemical Industry [source] Image analysis of Daphnia populations: non-destructive determination of demography and biomass in culturesFRESHWATER BIOLOGY, Issue 10 2002Per J. Færøvig SUMMARY 1. An image analysis technique was developed for the semiautomatic determination of abundance, size distribution and biomass in Daphnia cultures. This allowed detailed observations of growth, demography and biomass accumulation in live populations, avoiding artifacts caused by subsampling and sampling losses. 2. The image analysis method gave fast, non-destructive and reliable individual counts, even in cultures with high density and a large fraction of juveniles. 3. In Daphnia, animal width changes with nutritional status and growth within instar, while length changes only at the moult. Thus, estimation of individual biomass using an ellipsoidal model based on animal width gave improved biomass calculations compared to manual counting, sizing, and length : weight regressions. 4. The power of the image analysis technique for assessing population growth and size structure was demonstrated in two 40-day experiments, with Daphnia magna feeding on the green algae Selenastrum capricornutum in a two-stage chemostat system. [source] Morphometric relationship of length,weight and chelae length,width of eastern white river crayfish (Procambarus acutus acutus, Girard, 1852), under culture conditionsJOURNAL OF APPLIED ICHTHYOLOGY, Issue 5 2007Y. Mazlum Summary Length,weight (TL vs WWT) and chelae length,width (ChL vs ChW) relationships were described for juveniles, males and females, and for form I and form II males of Procambarus acutus acutus. The length,weight relationships for juveniles, form I, form II males, and females could be described as: WWT = 5 × 10,3 TL3.09, WWT = 6 × 10,3 TL3.61, WWT = 6 × 10,9 TL3.26, and WWT = 6 × 10,4 TL3.5, respectively. In all forms, growth was allometric (P < 0.05). The ancova test indicated that slopes and intercepts of the length,weight regressions were significantly different between sex and sexual stages. The regressions for chelae length,width relationships for form I and form II males, and females were: ChW = ,0.81 + 0.27CL, ChW = ,0.33 + 0.25CL, and ChW = ,0.82 + 0.32CL, respectively. Although the slope and intercepts of regressions for ChL and ChW were similar for those of form I and form II males, the slopes and intercepts of regressions of females were significantly different from form I and form II males. No statistical difference was observed in mean ChL between form II males and females (P > 0.05), but a significant difference was detected in mean ChL between form I and form II males (P < 0.05) and form I and females (P < 0.05). Form I males had longer ChL than form II males and females. The same trend was observed in mean ChW for form I and form II males, but a significant difference was detected between form II males and females (P < 0.05). In addition, results indicated that chelae lengths and widths increased allometrically with total length (TL) for both sex and sexual stages. [source] Age, growth and mortality of picarel, Spicara smaris L. (Pisces: Centracanthidae), from the eastern Adriatic (Croatian coast)JOURNAL OF APPLIED ICHTHYOLOGY, Issue 1 2003J. Dul Age, growth and mortality of picarel Spicara smaris L. collected in the eastern middle Adriatic during April and May in 1999 were studied. The total lengths of individuals ranged from 6.3 to 19.8 cm, and the weight from 2.2 to 78.3 g. The slopes (b values) of the total length,weight regressions indicated negative allometric growth for females (b=2.8± 0.03) and isometric growth for males (b=3.08±0.04). The oldest female and male were 4 and 6 years, respectively. The von Bertalanffy growth parameters for the sexes combined were (L,=22.76 cm, K=0.277, t0=,0.793). The overall sex ratio was 1 : 4.1 in favour of females. Individuals larger than 17.9 cm were all males. [source] Intrabasin variations in age and growth of Barbus bocagei populationsJOURNAL OF APPLIED ICHTHYOLOGY, Issue 3 2002J. M. Oliveira The age and growth of the Iberian cyprinid Barbus bocagei Steindachner 1865, were studied in six different streams in the River Tagus basin. The fish were aged by scale reading and their lengths were back-calculated. Age structures were similar at four sites (R. Ocreza, R. Rio Frio, R. Sever and R. Canha), where six to seven age classes were present. In contrast, 10 age groups were detected in the R. Erges and 11 in the R. Nabão. The slopes (b) of the length,weight regressions did not differ significantly between the six populations. The growth of the populations was compared with that available throughout the study area (another two sites). Classification analysis of the lengths-at-age by sites matrix revealed three major groups (by decreasing order of growth rate): 1, Belver Reservoir; 2, R. Tagus, Nabão and Erges; 3, Rio Frio, Sever, Canha and Ocreza. We hypothesize that highly stressful conditions in the summer diminish growth rates and longevity in the Rio Frio, Canha and Ocreza rivers and that lower temperatures all year round could limit growth in the Sever. Higher growth rates were observed in warm-water streams with more stable environmental conditions (group 2) and in a highly productive lentic system (group 1). Among the sites that were studied, the one that revealed the highest growth performance by B. bocagei was the most polluted, the R. Nabão. This offers evidence of the species plasticity and of its successful adaptation to streams displaying significant degradation in water quality. [source] Generalised regressions provide good estimates of insect and spider biomass in the monsoonal tropics of AustraliaAUSTRALIAN JOURNAL OF ENTOMOLOGY, Issue 3 2006Christopher J Brady Abstract, The estimation of arthropod biomass is often important in studies of terrestrial ecosystem structure and function, including analyses of the relative importance of different arthropod taxa in the diet of insectivorous animals. In order to estimate arthropod biomass in eucalypt woodlands and rehabilitated mine-land in the monsoonal tropics of northern Australia, insect morpho-species (n = 693) and spider morpho-species (n = 100) were collected, sorted, then weighed and measured. Body length,weight regressions were determined for spiders, nine insect orders and all insects combined. There was a significant relationship between body length and weight for all taxonomic groups, with the power model being a better predictor than linear or exponential models for all groups except Diptera (which was best predicted by the linear model). Whilst Schoener (1980) found that the length,weight regression slopes of neotropical insects (all orders combined, as well as several individual orders) differed from those of their temperate North American counterparts, our comparisons between monsoon-tropical and temperate Australian arthropods suggested differences among Dipterans and spiders only. We conclude that generalised regressions provide adequate estimates of arthropod biomass across Australia, providing that the body proportions of the dominant taxa do not vary substantively. [source] | |||||||||||||