Selection of durum wheat genotypes with wide adaptability across various environments is important before recommending them to reach a high rate of genotype adoption. Multi-environment grain yield trials of 20 durum wheat genotypes were conducted at five locations of Iran (Gachsaran, Gonbad, Moghan, Ilam and Khorram abad) over four years (2009-2013). Combined ANOVA of yield data of the 20 environments revealed highly significant differences among genotypes and environments as well as significant GE interaction indicated differential performance of genotypes over test environments. Results of F Ratio indicated that only five interaction principal components (IPCs) were significant at the 0.01 probability level. Also, the GE interaction is comprised of 29.7% noise and 70.03% signal. According to these distinct numbers of significant axes, fourteen AMMI stability parameters were computed. Finally according to the most of type 1 of AMMI parameters (EV1, AMGE1, SIPC1 and D1), genotypes G8, G17 and G11; based on the type 2 of AMMI parameters and ASV, genotypes G4, G5, G10, G11 and G17; due to type 3 of AMMI parameters and MASV, genotypes G8, G10 and G12 were detected as the most stable genotypes. Considering all of the AMMI stability parameters, genotypes G8, G10, G11, G12 and G17 following to genotypes G7 and G9 were the most stable genotypes. The best recommended genotypes according to the present study are G10 with 3470 kg ha-1 grain yield for Gachsaran and Khorramabad, G12 with 3343 kg ha-1 grain yield for Ilam and G10 and G12 for Moghan and Gonbad regions wich had high mean yield and were most stable for related mega-environments.

Genotype × environment interactions complicate selection of superior genotypes for narrow and wide adaptation. Multienvironment yield trials of twenty durum wheat genotypes were conducted at five locations of Iran (Gachsaran, Gonbad, Moghan, Ilam and Khorram abad) over four years (2009-2013). Combined ANOVA of yield data of the twenty environments (year/location combined) revealed highly significant differences among genotypes and environments as well as significant genotype-environment interaction indicated differential performance of genotypes over test environments. The GE interaction was examined using multivariate analysis technique as principal coordinate analysis (PCOA). According to grand means and total mean yield, test environments were grouped into two main groups as high mean yield (H) and low mean yield (L). There were eleven H test environments and nine L test environments which analyzed in the sequential cycles. For each cycle, both scatter point diagram and minimum spanning tree plot were drawn. The identified most stable genotypes with dynamic stability concept and based on the minimum spanning tree plots and centroid distances were G12 (3342 kg ha-1), G10 (3470.3 kg ha-1), G5 (3203.0 kg ha-1), and G1 (3263.5 kg ha-1), and therefore could be recommended for unfavorable or poor conditions. Genotypes G10 (3470.3 kg ha-1) and G9 (3404.2 kg ha-1) were located several times in the vertex positions of high cycles according to the principal coordinates analysis (PCOA) and therefore could be recommended for favorable or rich conditions. Finally, the results of principal coordinates analysis in general confirmed the breeding value of the genotypes, obtained on the basis of the yield stability evaluation.

The motivation behind this paper is to investigate the use of Softmax model for classification. We show that Softmax model is a nonlinear generalization for the logistic discrimination, that can approximate the posterior probabilities of classes where other Artificial neural network (ANN) models don't have this ability. We show that Softmax model has more flexibility than logistic discrimination in terms of correct classification. To show the performance of Softmax model a medical data set on thyroid gland state is used. The result is that Softmax model may suffer from overfitting.

No abstract