Performance Analysis and Validation of Fuzzy based Secure LAR Routing Protocol in MANETs using NN Tool

Artificial neural network can be used for predicting because of having the capability of examining and determining the historical data used for prediction. ANN has better accuracy than statistical and mathematical models. ANN works on the principle of biological neurons which is a type of data driven technique. The training simulation results indicate the feed forward back propagation model two input neurons and one output unit produce best predictive results for neural network. By varying the simulation, train the neural network in the performance function namely LAR, Secure LAR and FZ-S-LAR, the transfer functions TRANsig to validate the results. The results obtained after training the data under different scenarios mse gives better validation results. The errors obtained the training errors for mse is 0.99571, testing error is 0.98694 and validation is 0.98496.