Utilizing CLONALPropagation Algorithm for Pattern Matching with Training Data Sets

Pattern recognition has been around for many years in the world of computer engineering and science. It’s a technique of learning and matching pattern data sets through memorization. There are two main techniques being used typically by most models which are the back propagation and feed-forward network from ANN (Artificial Neural Network). Both are used simultaneously in completing a task given. Discussed together in this paper are also the features of CLONALNet which is a part of AIS (Artificial Immune System). CLONALNet is a technique derived from the clonal se- lection family where it is a hybrid design of CLONALG and opt-AINet used for optimization and hypermutation. Both techniques are then combined producing an- other hybrid algorithm known as CLONALPropagation build specifically for pattern recognition. Several parameters are chosen for hybrid purpose in this work. For ANN the main parameters chosen are the recognition rate, mse and β. As for AIS, the chosen parameters are α (mutation rate) and random population groups, Ab. The outcome of this research will be an algorithm which has a higher recognition rate with a lower or zero error rates, mse compared to the one in ANN back propagation technique.