A Graph-Based Method for Active Outlier Detection With Limited Expert Feedback

Abstract

Labeled data, particularly for the outlier class, are difficult to obtain. Thus, outlier detection is typically regarded as an unsupervised learning problem. However, it still has an opportunity to obtain few labeled data. For example, a human analyst can give feedback to few data when he/she examines the results of an unsupervised outlier detection method. Moreover, the widely used unsupervised method for outlier detection cannot only take the labeled data into consideration nor use them properly. In this study, we first propose a graph-based method to endow the unsupervised method with the ability to consider few labeled data. Then, we extend our semi-supervised method to active outlier detection by incorporating the query strategy that selects top-ranked outliers. Comprehensive experiments on 12 real-world datasets demonstrate that our semi-supervised outlier detection method is comparable with the best of state-of-the-art approaches, and our active outlier detection method outperforms state-of-the-art methods.

An Enhanced Outlier Detection Approach for Multidimensional Datasets Using a Synergistic Firefly and Grey Wolf Optimization-Based Method

“…Li et al [23] introduced a graph-based strategy, enabling an unsupervised approach to evaluate a minimal amount of labeled data. The semi-supervised system was extended to active outlier detection by incorporating a query method that selected top-ranked outliers.…”

A rapid outliers detection and correction method for external ballistic velocity measurement data

“…Otherwise, a novel single-objective generative adversarial active learning method was proposed to enrich the information on outliers [24]. And to detect outliers in a few labeled data, a graph-based semisupervised method was designed [25]. However, deep learning models often require a large amount of training data, so they cannot be applied to a small sample of data.…”