BazTech - Yadda

Ograniczanie wyników

Znaleziono wyników: 2

Liczba wyników na stronie

Wyniki wyszukiwania

Wyszukiwano:
w słowach kluczowych: globalna analiza wrażliwości

Sortuj według:

Ogranicz wyniki do:

Parameter sensitivity and uncertainty of radiation interception models for intercropping system

Zeng Wenzhi, Lu Yuchao, Srivastava Amit Kumar, Gaiser Thomas, Huang Jiesheng

Ecological Chemistry and Engineering. S

2020

Vol. 27, nr 3

437--456

Estimating the interception of radiation is the first and crucial step for the prediction of production for intercropping systems. Determining the relative importance of radiation interception models to the specific outputs could assist in developing suitable model structures, which fit to the theory of light interception and promote model improvements. Assuming an intercropping system with a taller and a shorter crop, a variance-based global sensitivity analysis (EFAST) was applied to three radiation interception models (M1, M2 and M3). The sensitivity indices including main (Si) and total effects (STi) of the fraction of intercepted radiation by the taller (ftaller), the shorter (fshorter) and both intercrops together (fall) were quantified with different perturbations of the geometric arrangement of the crops (10-60 %). We found both ftaller and fshorter in M1 are most sensitive to the leaf area index of the taller crop (LAItaller). In M2, based on the main effects, the leaf area index of the shorter crop (LAIshorter) replaces LAItaller and becomes the most sensitive parameter for fshorter when the perturbations of widths of taller and shorter crops (Wtaller and Wshorter) become 40 % and larger. Furthermore, in M3, ftaller is most sensitive to LAItaller while fshorter is most sensitive to LAIshorter before the perturbations of geometry parameters becoming larger than 50 %. Meanwhile, LAItaller, LAIshorter, and Ktaller are the three most sensitive parameters for fall in all three models. From the results we conclude that M3 is the most plausible radiation interception model among the three models.

Application of neural networks for social capital analysis

Siderska J.

Czasopismo Techniczne. Automatyka

2013

R. 110, z. 2-AC

57--66

The paper investigates the possibility of using soft computing for estimating the value of social capital. Our approach is applied to the case of Red Hat Inc. – the world’s leading provider of open source solutions. The objective of the research was to develop an artificial neural network for forecasting the value of social capital. These studies also allow us to identify variables significantly affecting the value of social capital. Computer simulations and assessments were done using software package STATISTICA Automated Neural Networks. The paper concludes with discussion and proposals for further research.

Głównym celem artykułu jest analiza możliwości zastosowania obliczeń inteligentnych do modelowania kapitału społecznego firmy Red Hat Inc. – światowego lidera rozwiązań open source dla biznesu. Zasadniczym celem badań jest zaproponowanie struktury sztucznej sieci neuronowej do analizy wartości kapitału społecznego. Zidentyfikowano zmienne istotnie wpływające na wartość tego kapitału. Wszystkie symulacje komputerowe oraz oszacowania przeprowadzono w pakiecie statystycznym STATISTICA Automatyczne Sieci Neuronowe. W artykule przedyskutowano wyniki testów otrzymanych z zastosowaniem zaprojektowanego modelu oraz zaproponowano tematykę dalszych badań.