Abstract
Large number of outlier detection methods have emerged in recent years due to their importance in many real-world applications. The graph-based methods, which can effectively capture the inter-dependencies of related objects, is one of the most powerful methods in this area. However, most of the graph-based methods ignore the local information around each node, which leads to a high false-positive rate for outlier detection. In this study, we present a new outlier detection model, which combines the graph representation with the local information around each object to construct a local information graph, and calculates the outlier score by performing a random walk process on the graph. Local information graph is constructed to capture the asymmetric inter-dependencies relationship between various types of data objects. Based on two different types of restart vectors to solve the dangling link problem, we propose two distinct algorithms for outlier detection. Experiments on synthetic datasets indicate that the proposed algorithms could efficiently distinguish outlier objects in different distributed datasets. Furthermore, the results on a number of real-world datasets also show that our approaches outperform the state-of-the-art outlier detection methods.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.