Naslov
The valuation of anthropogenic soils using neural network

Autori
Knezić, Snježana ; Miloš, Boško

Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni

Izvornik
Proceedings of ITI '94 / Ceric, V. ; Hljuz Dobric, V. - Zagreb : Univ. Computing Centre, 1994, 125-130

Skup
16th International Conference `Information Technology Interfaces'

Mjesto i datum
Pula, Hrvatska, 14.06.1994. - 17.06.1994

Vrsta sudjelovanja
Predavanje

Vrsta recenzije
Međunarodna recenzija

Ključne riječi
anthropogenic soils; soil valuation; neural network; soil suitability classification; soil properties; physiological plant demand; semantic estimation; calciphobic fruit-trees; training set; fuzzy set theory; tree growing; knowledge base; pedological decision support system; horticulture

Sažetak
Describes the application of a neural network for the evaluation of anthropogenic soils. Soil suitability is defined as a function of the compatibility between the soil properties and physiological plant demand, and it is given by semantic estimation. In such a classification, it is not possible to define rigid class limits, so the principle of fuzzy set theory is applied. Since the training set represents a model for classification (soil valuation), it is very important to make it consistent and well-structured. Testing and application of a trained neural network for soil suitability classification for growing calciphobic fruit-trees showed that this method is attractive for creating a knowledge base as a part of a pedological decision support system.

Izvorni jezik
Engleski

Znanstvena područja
Temeljne tehničke znanosti

POVEZANOST RADA