Attribute weighting via differential evolution for attribute weighted clonal selection algorithm

Classification is an important technology in data mining, while clonal selection algorithm (CSA) is a very effective classification method. Although CSA brings a new effective tool for solving complex problems, we can not completely say that it over-performs to other algorithms especially in the classification field. A main problem of CSA classifier is that it does not carry attribute imbalance. It uses a pure distance criterion to calculate affinity degree of the antibody and antigen. So we utilize weighting attribute scheme to balance the effects of attributes in classification process and attribute weighted CSA (AWCSA) comes into existence. The efficiency of AWCSA lies mainly in the attribute weighting scheme it uses. In this paper we use differential evolution (DE) algorithm to determine the weights of attributes and then use these weights in AWCSA. We evaluate the performance of new algorithm (DE-AWCSA) on six standard datasets. Experimental results show that this attribute weighting process highly benefits the classification accuracy