Date of Publication :7th September 2017

Abstract: Outlier detection has been used to detect the outlier and, where appropriate, eliminate outliers from various types of data. It has vital applications in the field of fraud detection, network robustness analysis, Insider Trading Detection, email spam detection, Medical and Public Health Outlier Detection, Industrial Damage Detection, Image processing fraud detection, marketing, network sensors and intrusion detection. In this paper, we propose a kmean clustering and neural network as novel to detect the outlier in network analysis. Especially in a social network, k means clustering and neural network is used to find the community overlapped user in the network as well as it finds more kclique which describe the strong coupling of data. In this paper, we propose that this method is efficient to find out outlier in social network analyses. Moreover, we show the effectiveness of this new method using the experiments data. (Abstract)

Reference :