Explainable multi-class anomaly detection on functional data

  title={Explainable multi-class anomaly detection on functional data},
  author={Mathieu Cura and Katarina Firdova and C{\'e}line Labart and Arthur Martel},
. 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. 



Explainable anomaly detection on high-dimensional time series data

This paper presents an approach to integrated anomaly detection (AD) and explanation discovery (ED), which is able to leverage state-of-the-art Deep Learning techniques for anomaly detection, while being able to recover human-readable explanations for detected anomalies.

Exathlon: A Benchmark for Explainable Anomaly Detection over Time Series

This paper presents Exathlon, the first comprehensive public benchmark for explainable anomaly detection over high-dimensional time series data and demonstrates the practical utility of Exathlon’s dataset, evaluation methodology, and end-to-end data science pipeline design through an experimental study with three state-of-the-art AD and ED techniques.

An Anomaly Detection and Explainability Framework using Convolutional Autoencoders for Data Storage Systems

An endto-end framework based on convolutional autoencoders which not only allows for anomaly detection on multivariate time series data, but also provides explainability, which enables support engineers to effectively tackle abnormal behaviors, all while allowing for crucial feedback.

Explainable k-Means and k-Medians Clustering

It is shown that popular top-down decision tree algorithms may lead to clusterings with arbitrarily large cost, and that any tree-induced clustering must in general incur an $\Omega(\log k)$ approximation factor compared to the optimal clustering.

Outlier detection in functional data by depth measures, with application to identify abnormal NOx levels

This paper analyzes outlier detection for functional data by means of functional depths, which measures the centrality of a given curve within a group of trajectories providing center‐outward

Functional Isolation Forest

This paper addresses the issue of (randomly) splitting the functional space in a flexible manner in order to isolate progressively any trajectory from the others, a key ingredient to the efficiency of the algorithm.

Functional Data Analysis

  • H. Müller
  • Computer Science
    International Encyclopedia of Statistical Science
  • 2011
An overview of functional data analysis is provided, starting with simple statistical notions such as mean and covariance functions, then covering some core techniques, the most popular of which is functional principal component analysis (FPCA).

Beyond Sparsity: Tree Regularization of Deep Models for Interpretability

This work explicitly regularizes deep models so human users might step through the process behind their predictions in little time, and trains deep time-series models so their class-probability predictions have high accuracy while being closely modeled by decision trees with few nodes.

A Unified Approach to Interpreting Model Predictions

A unified framework for interpreting predictions, SHAP (SHapley Additive exPlanations), which unifies six existing methods and presents new methods that show improved computational performance and/or better consistency with human intuition than previous approaches.