Forecasting Tree Volume Growth Using Artificial Neural Networks The Case Of Cypresses Lusitania Species

Neural networks are computational models with the capacity tornlearn or generalize complex relationships that exist among data.rnAlthough there are different kinds of networks, the multi-layeredrnfeed-forward neural network is the most widely used network thatrnis capable of representing non-linear functional mappings betweenrninputs and outputs. The training of this network is accomplishedrnby the method of error back-propagation In this paper, the feed-forward neural network with backrnpropagation learning algorithm is presented for forecasting treernvolume growth of Cupressus lustanica species. The data set forrntraining the neural network was obtained from the ForestryrnResearch Center (FRC). As an input to the neural network, thernhistorical tree volumes are used to train the network. After trainingrnthe network, the results of forecasting is evaluated using test datarnset. The result indicates that the model yields good prediction withrnindependent test data set, providing about 86.7% correct forecastsrnwithin ±2cm3 of the observed values. This suggests that the neuralrnnetwork is a good candidate for forecasting future value of treernvolume given properly and accurately measured historical data.

Get Full Work

Report copyright infringement or plagiarism