Multiscale Comparison of Nonparametric Trend Curves
Multiscale Comparison of Nonparametric Trend Curves
Authors: Marina Khismatullina, Michael Vogt Date: 2026-06-01 Paper ID: openalex:2209.10841
Summary
This paper proposes a rigorous econometric framework for the comparative analysis of nonparametric time trends across multiple series. The methodology employs a multiscale testing approach, which provides statistically sound confidence statements regarding whether trends differ and precisely which time intervals contain significant differences. By leveraging these tests, the authors derive an automated clustering algorithm that groups time series based on shared trend characteristics, supported by asymptotic theory and empirical applications.
Key Contributions
- Develops a multiscale testing framework for comparing nonparametric trends in multiple time series.
- Enables precise identification of time intervals where specific trends significantly deviate.
- Introduces a clustering algorithm based on the proposed multiscale test to categorize time series by trend similarity.
Links
Metadata & Links
- url
- https://arxiv.org/abs/2209.10841
- paper_id
- 2209.10841
- paper_source
- openalex
- domain
- time-series
- tags
-
- architectures
-
- datasets
-
- concept_slugs
-
- dataset_slugs
-
- skill
- TimeSeriesSkill
- created_at
- 2026-06-04T06:34:27Z