Enhancing Outlier Detection by an Outlier Indicator

Abstract

Outlier detection is an important task in data mining and has high practical value in numerous applications such as astronomical observation, text detection, fraud detection and so on. At present, a large number of popular outlier detection algorithms are available, including distribution-based, distance-based, density-based, and clustering-based approaches and so on. However, traditional outlier detection algorithms face some challenges. For one example, most distance-based and density-based outlier detection methods are based on k-nearest neighbors and therefore, are very sensitive to the value of k. For another example, some methods can only detect global outliers, but fail to detect local outliers. Last but not the least, most outlier detection algorithms do not accurately distinguish between boundary points and outliers. To partially solve these problems, in this paper, we propose to augment some boundary indicators to classical outlier detection algorithms. Experiments performed on both synthetic and real data sets demonstrate the efficacy of enhanced outlier detection algorithms.