Skip to Main Content
Cognitive maps, one of the hot topic in the research of computational intelligence, have been widely used in knowledge representation and decision-making. In mining of cognitive maps on the basis of data resources, outlier data seriously affect the accuracy of cognitive maps. Therefore, this paper, based on the analysis of traditional ones, proposes a new outlier data detection algorithm. The algorithm firstly partitions the entire data set with the hierarchical clustering algorithm, then rules out the partitions that do not contain abnormal data, and finally detects outlier data in the remaining partitions. Experimental results show that the algorithm, compared with the traditional ones, reduces the required amount of the computer memory and enhances efficiency.