期刊名称：International Journal of Advances in Soft Computing and Its Applications
印刷版ISSN：2074-8523
出版年度：2014
卷号：6
期号：3
出版社：International Center for Scientific Research and Studies
摘要：The dendritic cell algorithm is an effective technique to detect anomalies in time series applications. However, the algorithm is less effective when it mines a general classification dataset because the items are not organized in an orderly event-driven manner. Ideally, for they need to be arranged in sequence by sorting them according to decision class. However, it is not practicable to apply this step because the decision classes for real datasets is unknown. Therefore, an integrated model that combines the dendritic cell algorithm and the k-means algorithm is proposed as an alternative to the existing sorting function based on decision class. The proposed model is evaluated by applying it to eight universal classification datasets and assessing its performance according to four evaluation metrics: detection rate, specificity, false detection rate, and accuracy. The results show that the proposed clustered dendritic cell algorithm is more effective than the non-clustered version. When applied to a benchmark dataset, the clustered dendritic cell algorithm demonstrates significant improvement in performance on the unordered version of the dataset and generates a comparable result to that of its competitor. For the other seven datasets, the proposed algorithm generates better specificity, false detection rate, and accuracy. The findings indicate that item–centroid distance within a cluster can be adopted to transform an unordered dataset into a sequential dataset, thus fulfilling the dendritic cell algorithm requirement for ordered data
关键词：Artificial immune system; clustering; dendritic cell algorithm; k- ; means
