Nature and Science

Applying Self-Organizing Competition Artificial Neural Networks to Classify the Soil

Xiangdong Ma*, Qiang Fu**

Abstarc :

Through applying the clustering function of Self-Organizing Mapping (SOM) network, the writer uses MATLAB 5.3 software to classify 21 kinds of soil samples in Sanjiang Plain. Through comparing the result with references that uses the method of fuzzy clustering, the paper concludes that the SOM network can reflect the complicated information among each soil samples. The effect of classification is good, and can be applied on soil classification. [Nature and Science 2003;1(1):75-81].

Key Word :
SOM network; soil classification

Volume 1, Number 1, November 2003 , ISSN 1545-0740