Keywords: Missing Data, Time Series, Imputation, Visualization, Data Pre-Processing
Webpages:
https://CRAN.R-project.org/package=imputeTS,
https://github.com/SteffenMoritz/imputeTS In almost every domain from industry, finance, up to biology time series data is measured. One common problem that can come along with time series measurement are missing observations. During several projects with industry partners in the last years, we often experienced sensor malfunctions or transmission issues leading to missing sensor data. As subsequent processes or analysis methods may require missing values to be replaced with reasonable values up-front, missing data handling can be crucial.
This talk gives a short overview about methods for missing data in time series in
R in general and subsequently introduces the
imputeTS package. The
imputeTS package is specifically made for handling missing data in time series and offers several functions for visualization and replacement (imputation) of missing data. Based on usage examples it is shown how
imputeTS can be used for time series imputation.
Most well-known and established packages (e.g.
mice,
VIM,
AMELIA,
missMDA) for missing value imputation focus mostly on cross-sectional data, while methods for time series data are not that familiar to users. Also, from an algorithmic perspective, these two imputation use cases are slightly different: imputation for cross-sectional data relies on inter-attribute correlations, while (univariate) time series imputation needs to employ time dependencies. Overall, this talk is supposed to give users a first glance at time series imputation in
R with special focus on the
imputeTS package.