Ancillary techniques for neural network applications

To a large extent, the successful implementation of neural nets depends on several ancillary techniques for data preprocessing, training and testing. Some of these techniques are investigated and discussed in this paper. They include genetic algorithm, fuzzy logic theory, query-based learning and feature extraction. For each technique, the paradigm, theory and application are described. The advantages of the application of these ancillary techniques for the neural networks are also listed. The simulation results for each proposed technique showed their significant role and practicality