Showing 4 results for Discriminant Analysis
A. Dehdari, A. Rezai, M. Mobli,
Volume 5, Issue 2 (7-2001)
Abstract
In order to evaluate the morphological and agronomic characteristics of 19 land races of onions and one foreign cultivar (Yellow Sweet Spanish), an experiment was conducted in 1998 at the Research Farm of the College of Agriculture, Isfahan University of Technology.
Analysis of variance showed significant differences among genotypes for all of the morphological and agronomic traits. Dry weight and number of days to emergence had the highest and lowest coefficients of genetic variability, respectively. Plant fresh weight, yield of 30 plants and total yield also had high coefficients of variability. Broad sense heritability estimates were high for all of the traits, indicating low environmental affects them. Based on cluster analysis and plot of the first two canonical variables, the genotypes were classified in four groups with different agronomic traits. Canonical discriminant analysis based on nine agronomic traits introduced three canonical variables which justified 99.9 percent of the total variation among characters. Principal component analysis revealed four components while factor analysis showed three factors which explained 87.3 and 95.5 percent of the total variation among characters, respectively. The first and second factors were related to adaptation and assimilate translocation, respectively. Bulb diameter, bulb height and bulb weight in positive directions and sensitivity to Fusarium in negative direction had greater loads in the third factor.
J. Mohammadi, H. Khademi, M. Nael,
Volume 9, Issue 3 (10-2005)
Abstract
In order to achieve a sustainable management of land resources and to improve land quality, quantitative assessment of effective factors and soil quality indicators are required. The aim of this study was to evaluate variability of selected soil quality attributes in central Zagros affected by such factors as region, land use and management practices. Twelve sites were selected in three provinces including Chahar Mahal va Backtiari (Sabzku, Broujen), Isfahan (Semirum), and Kohkeloyeh va Boyerahmad (Yasodje). Different management practices were considered such as: protected pasture, intensive grazing, controlled grazing, dryland farming, irrigated wheat cultivation, legume-farming practice, protected forest, and degraded forest. Systematic sampling with taking 50 samples of surface soil in each site was carried out. The results of univariate and multivariate analysis revealed that all factors significantly influenced the spatial variability of selected soil quality attributes namely phosphatase activity, microbial respiration, soil organic matter, and total nitrogen. The results obtained from discriminant analysis indicated that all selected soil quality parameters could significantly be used as soil quality indicators in order to recognize and discriminate sustainable agricultural and forestry ecosystems and/or optimal management practices.
E. Karami, K. Rezaei- Moghaddam, H. Ebrahimi,
Volume 10, Issue 1 (4-2006)
Abstract
Increasing the water use efficiency through promotion of sprinkler irrigation system, in order to increase production, has been the aim of the Ministry of Agricultural-Jihad in the past decade. Considerable amount of investment and efforts have been devoted to this end. Therefore, investigation of these efforts and development of models to predict the adoption behavior of farmers are of considerable importance. A survey research was conducted in four provinces (Fars, Bushehr, Kohkiloye-va-Boyerahmad, and Chaharmahal-va-Bakhteyari). A stratified random sample of 422 farmers was interviewed including 124 and 298 adopters and non-adopters, respectively. The findings indicated that the discriminant model developed based on the multiplicity model is a better predictor of farmers’ adoption behavior than diffusion and farm structure model. Application of multiplicity model in extension of sprinkler irrigation can increase the adoption rate and as result the efficiency of extension efforts.
W. Zarei, M. Sheklabadi,
Volume 18, Issue 70 (3-2015)
Abstract
The aim of the study was to investigate the effects of land use on soil quality parameters using multivariate statistical analysis. Soil samples (0-25 and 25-50 cm depths) were taken from three land uses in forest area of Marivan including forest, rangeland, and cultivated land. Soil characteristics of pH, EC, sand, silt, clay and CaCO3 content, water-stable aggregates and their organic carbon content were measured. Principal component, cluster and discriminant analyses were used to evaluate the soil quality. Principal component analysis classified soil properties into five factors. The most important factors were soil aggregates organic carbon content and aggregate stability indices. Schematic distribution of factors and also cluster analysis showed the same pattern. Soil aggregates organic carbon content, water-stable aggregates and aggregate stability indices were the most sensitive factors to land use changes. These soil properties and factors had the same pattern in forest and rangeland, but significantly reduced in the cultivated land use. Land use change from forest to cultivated land resulted in significant decrease of aggregates organic carbon content, water-stable aggregates and also an increase of pH. The results showed the usefulness of multivariate statistical methods for integration of the soil properties and determination of different soil quality indices.
