Effect of Fuzzy Resource Allocation Method on Airs Classifier Accuracy

Artificial Immune Recognition System (AIRS) is an immune inspired classifier that is comparable to many popular classifiers. Many researches have been conducted to improve the accuracy of AIRS and to identify the significant components of AIRS that could empower it for better performance. Some of these researches have focused on the resource allocation component of AIRS. This study investigates the difference between the accuracy of AIRS using a fuzzy resource allocation approach with the accuracy of the current resource allocation technique by using statistical methods. The combination of ten-fold cross validation and t-test was used as evaluation method and algorithms tested on benchmark datasets of UCI machine learning repository. Based on the results of experiments, using the fuzzy resource allocation technique increases the accuracy of AIRS in majority of the datasets. However, the increase is significant in minority of datasets.