Explainable multi-class anomaly detection on functional data

TL;DR

This paper presents an anomaly detection method for multivariate functional data that combines feature transformation, Isolation forest, and explainability via SHAP coefficients and decision trees, validated on simulated and industrial data.

Contribution

It introduces a novel approach integrating anomaly detection with explainability specifically for multivariate functional data.

Findings

Effective anomaly detection demonstrated on simulated data

Successful application to real industrial data

Explainability method provides interpretable insights

Abstract

In this paper we describe an approach for anomaly detection and its explainability in multivariate functional data. The anomaly detection procedure consists of transforming the series into a vector of features and using an Isolation forest algorithm. The explainable procedure is based on the computation of the SHAP coefficients and on the use of a supervised decision tree. We apply it on simulated data to measure the performance of our method and on real data coming from industry.