In recent years neural computing has emerged as a practical technology, with successful applications in many fields. The use of neural networks in pattern classification is becoming increasingly widespread, with applications in signal processing areas such as signal detection and classification. In this book, the signals concerned include sonar and radar ionosphere databases from the research literature. These two data sets are intentionally chosen, because they contain high dimensionality, small sample sized problem and complex decision boundaries due to overlapping clusters. Learning from small sample sized dataset is typically a very difficult problem in the theory of complexity. It is a challenging task even for neural network. We have investigated the neural network based design of an optimal classifier and attempt is made to suggest suitable model by comparative analysis of the designed classifier for pattern classification on standard benchmark databases of sonar and radar ionosphere from the real world systems.