Classification is one of the matters discussed in the field of data mining, being of high importance in knowledge discovery. Many data mining techniques have been applied to deal with classification problems. Decision trees are one of the oldest and most well-known methods to construct classification models, in a way that the model implements itself. In this paper, a method which is the combination of the powerful and evolutionary algorithm of GP and the Instance-based learning algorithm of KNN is presented, acting efficiently in separating classes and data. This method is proposed in order to check and improve the presented method of MGP (Multi-age Genetic Programming) and to construct decision trees. The simulation of the proposed method has been done using the software MATLAB on the data collection of UCI. The results indicate an improvement in classification accuracy and a reduction in the runtime of the algorithm compared to the MGP and other methods.

Full text: IJEIR_1179_Final.pdf