Trivial Graph Features and Classical Learning are Enough to Detect Random Anomalies

TL;DR

This paper demonstrates that simple graph features combined with classical machine learning methods are highly effective for detecting random anomalies in link streams, offering a computationally efficient and interpretable solution.

Contribution

The study shows that trivial graph features and classical learning techniques outperform complex methods in detecting random anomalies in link streams.

Findings

Simple features and classical methods achieve high detection accuracy.

The approach is computationally efficient and easily interpretable.

Complex anomalies require more advanced detection techniques.

Abstract

Detecting anomalies in link streams that represent various kinds of interactions is an important research topic with crucial applications. Because of the lack of ground truth data, proposed methods are mostly evaluated through their ability to detect randomly injected links. In contrast with most proposed methods, that rely on complex approaches raising computational and/or interpretability issues, we show here that trivial graph features and classical learning techniques are sufficient to detect such anomalies extremely well. This basic approach has very low computational costs and it leads to easily interpretable results. It also has many other desirable properties that we study through an extensive set of experiments. We conclude that detection methods should now target more complex kinds of anomalies.