A vertical outlier detection algorithm with clusters as by-product

Abstract

Outlier detection can lead to discovering unexpected and interesting knowledge, which is critically important to some areas such as monitoring of criminal activities in electronic commerce, credit card fraud, and the like. In this paper, we propose an efficient outlier detection method with clusters as by-product, which works efficiently for large datasets. Our contributions are: a) We introduce a Local Connective Factor (LCF); b) Based on LCF, we propose an outlier detection method which can efficiently detect outliers and group data into clusters in a one-time process. Our method does not require the beforehand clustering process, which is the first step in other state-of-the-art clustering-based outlier detection methods; c) The performance of our method is further improved by means of a vertical data representation, Ptrees 1 . We tested our method with real dataset. Our method shows around five-time speed improvements compared to the other contemporary clustering-based outlier-detection approaches.

A method for outlier detection based on cluster analysis and visual expert criteria

“…There are a great many proximity‐based approaches. From the proximity viewpoint, an object is considered an outlier if it is not close to the other objects (Kollios, Gunopulos, Koudas, & Berchtold, ; Ren, Rahal, & Perrizo, ).…”

A New Group Key Management Structure for Fraudulent Internet Banking Payments Detection

“…This requires previous fraudulent use before any future fraud can be detected. Unsupervised detections does not rely on previous fraud cases but focuses on unusual transaction behaviour (Angiulli, F, 2006) ( Ren, D, 2004).…”

“…Better fraud detection has become an essential requirement for banks in order to maintain a viability payment system. At present, fraud detection is conducted using data mining, statistics, and artificial intelligence (Ghosh, S, 1994) ( Joris C, 2002) ( Ren, D, 2004). Such methods still lack sufficiently secure payment mechanisms to identify internal fraud and trace fraudulent transactions.…”

Why General Outlier Detection Techniques Do Not Suffice for Wireless Sensor Networks

“…introduce a new definition of a cluster-based local outlier, which takes size of a point's cluster and distance between the point and its closest cluster into account. Ren et al (2004) propose a more efficient clustering-based local outlier detection approach, which combines detection of outliers with grouping data into clusters in a one-time process. Bohm et al (2008) propose a robust clustering-based approach, which can be applied to a data set with non-Gaussian distribution to efficiently filter out the outliers.…”