predicting spatial distribution of redroot pigweed (amaranthus retroflexus l.) using the rbf neural network model

Estimating the spatial distribution of weeds for sitespecific control is essential. therefore, this research was conducted to predict and interpolate the spatial distribution of amaranthus retroflexus l. populations using a radial basis function neural network (rbfnn) in two potato fields. weed population data were collected from sampling 200 and 36 points, respectively, in two commercial potato fields in jolge rokh, of torbat heidarieh in khorasan razavi and mojen of shahroud in semnan provinces, iran, in 2012. some statistical tests, such as comparisons of the means, variance and statistical distribution, as well as linear regression, were used for the observed point sample data and the estimated weed seedling density surfaces to evaluate the neural network capability for predicting the spatial distribution of the weed. the results showed that the trained rbfnn had high capability in the spatial prediction in points that were not sampled with 100% output, 0.999 coefficients, and an average error of less than 0.04 and 0.07 in the mojen and jolge rokh regions, respectively. test results also showed that there was no significant difference between the statistical characteristics of actual data and the values predicted by the rbf-nn. according to the experimental results, the rbf-nn can be used as an alternative method to estimate the spatial changes function of annual weeds with random dispersion, such as redroot pigweed.