Problem
Research questions and friction points this paper is trying to address.
Automated detection of time series outliers
Modular pipeline construction for outlier analysis
Data-driven pipeline discovery for optimal detection
TODS: An Automated Time Series Outlier Detection System
🤖 AI Summary
To address the high deployment barrier and inflexible workflow customization in time-series anomaly detection, this paper proposes a scalable, primitive-based architecture specifically designed for time-series anomaly detection. The system introduces a modular primitive library comprising 70 composable base operators—covering preprocessing, feature analysis, detection algorithms, and refinement modules—enabling zero-code or low-code end-to-end pipeline construction. It integrates a GUI-based drag-and-drop interface with a data-driven automatic pipeline search algorithm, and adopts a Python microservice architecture to ensure high cohesion and loose coupling. The open-source framework (Apache 2.0 license, hosted on GitHub) bridges research reproducibility and industrial usability. Validated across multiple industry domains, it significantly lowers the engineering barrier for deploying time-series anomaly detection systems while maintaining flexibility and extensibility.
📝 Abstract
We present TODS, an automated Time Series Outlier Detection System for research and industrial applications. TODS is a highly modular system that supports easy pipeline construction. The basic building block of TODS is primitive, which is an implementation of a function with hyperparameters. TODS currently supports 70 primitives, including data processing, time series processing, feature analysis, detection algorithms, and a reinforcement module. Users can freely construct a pipeline using these primitives and perform end- to-end outlier detection with the constructed pipeline. TODS provides a Graphical User Interface (GUI), where users can flexibly design a pipeline with drag-and-drop. Moreover, a data-driven searcher is provided to automatically discover the most suitable pipelines given a dataset. TODS is released under Apache 2.0 license at https://github.com/datamllab/tods. A video is available on YouTube (https://youtu.be/JOtYxTclZgQ)
Innovation
Methods, ideas, or system contributions that make the work stand out.
Modular system with 70 customizable primitives
GUI for drag-and-drop pipeline construction
Data-driven searcher for optimal pipeline discovery
🔎 Similar Papers
No similar papers found.
