Analyzing a Large Number of Time Series : Learning and Monitoring Complex Patterns.DE (Statistics for Industry, Technology, and Engineering)

Kenett, Ron/ Ravishanker, Nalini/ Reis, Marco S.

Springer, Berlin; Birkhäuser（2026/07発売）

• 製本 Hardcover:ハードカバー版
• 商品コード 9783032335531

Description

The modern availability of multistream data in a wide range of applications is pushing the need to address time series data analysis in new and innovative ways. Sensor data in Industry 4.0 applications is one such example, and patient monitoring in emergency rooms is another. This context leads to challenges coming from unsynchronized data resolution, a variety of data structures, including imaging, and complex challenges in data integration.

This monograph will fill a gap in the methodology and practice of analyzing a large number of time series with complex temporal dependencies. It will focus on time series that have applications in disciplines such as business, climate, engineering, finance, health analytics, remote sensing, and manufacturing, providing novel and sophisticated data analytic techniques for researchers and practitioners. Specific topics that will be covered include state-of-the-art data science methods that enable systematic learning and monitoring of complex dependence patterns in a large number of time series; supervised and unsupervised learning (classification, clustering, biclustering) methods which provide monitoring and diagnostic capabilities to the analysis of time series; and monitoring and control of many time series in order to facilitate decision making.

1 - Introduction.- 2 - Time Series Analysis.- 2.1 - ARIMA Models.- 2.2 - Vector ARIMA Models.- 2.3 - Dynamic Linear Models.- 2.4 - High-Dimensional Dynamic Modeling.- 3 - Motivating Examples.- 3.1 - Multivariate Quality Control.- 3.2 - Industrial Process Monitoring.- 3.3 - Finance Applications.- 3.4 - Soft Sensor Development.- 4 - Supervised Learning for Time Series.- 4.1 - Classical Time Domain Methods.- 4.2 - Classical Frequency Domain Methods.- 4.3 - Tree-Based Methods.- 4.4 - Bayesian Methods.- 4.5 - Feature-Based Predictive Modeling.- 5 - Unsupervised Learning for Clustering Time Series.- 5.1 - Data-Based Clustering.- 5.2 - Feature-Based Clustering.- 5.3 - Model-Based Clustering.- 5.4 - Incorporating Domain Knowledge.- 6 - Biclustering.- 6.1 - Detection of Patterns.- 6.2 - Diagnostic Windows.- 6.3 - Clustering versus Biclustering.- 7 - Monitoring and Controlling Patterns.- 7.1 - Monitoring a Single Time Series.- 7.2 - Monitoring Many Independent Time Series.- 7.3 - Monitoring Multivariate Time Series.- 7.4 - Monitoring Multi-Stage Processes.- 7.5 - Multiscale Methods.- 7.6 - Process Monitoring of High-Dimensional Streams.- 8 - Review of Case Studies.