ISOLATED ARABIC OPTICAL CHARACTER RECOGNITION USING COMBINATION MODEL OF C4.5 AND GENETIC PROGRAMMING

Since recognizing of Arabic Characters by a junction is considered as one of the most important problems, this paper presents a combination system based on the evolutionary computation and machine learning algorithms due to solving it Since genetic programming, evolutionary computation algorithm, requires large resources to construct and evolve each generation, it s better to construct a near optimal initial population. As a result a machine learning algorithm is needed. C4.5, the machine learning algorithm, constructs the initial population based on its concept of decision tree with minimum description length This paper introduces a hybrid model of the machine learning algorithm C4.5 and genetic programming due to discovering the optimal local rules that classify isolated Arabic Characters. This new hybrid model achieves good results, since the short in genetic programming algorithm is neglected by the machine learning algorithm C4.5 and then the advantages of both techniques is integrated. The details of the new model are discussed whereas a comparative study is illustrated