R Package imputeTestbench to Compare Imputation Methods for Univariate Time Series

Area of Science:

• Statistics
• Data Science
• Time Series Analysis

Background:

• Missing observations are prevalent in time series datasets.
• The accuracy of imputation methods is influenced by data characteristics and missing data patterns.
• Existing imputation methods vary in complexity and effectiveness.

Purpose of the Study:

• To introduce the imputeTestbench package for comparing time series imputation methods.
• To assess how the amount and type of missing data affect imputation accuracy.
• To provide a flexible framework for evaluating and visualizing imputation algorithm performance.

Main Methods:

• The imputeTestbench package simulates missing data by random or block removal.
• It includes various imputation algorithms, from simple mean replacement to complex interpolation.
• Users can customize the package by adding or removing imputation methods.

Main Results:

• The package allows for comparative analysis of imputation method accuracy under different missing data scenarios.
• Visualizations highlight the performance differences between various imputation algorithms.
• Example applications demonstrate how dataset characteristics influence method effectiveness.

Conclusions:

• imputeTestbench offers a robust tool for understanding imputation method performance.
• It facilitates informed selection of imputation techniques based on specific data and missingness properties.
• The package supports enhanced accuracy in time series analysis by optimizing missing data handling.