Missing Data in Time Series: A Review of Imputation Methods and Case Study

Dealing with missingness in time series data is a very important, but oftentimes overlooked, step in data analysis. In this paper, the nature of time series data and missingness mechanisms are described to help identify which imputation method should be used to impute missing data, along with a review of imputation methods and how they work. Recommended methods from literature are used to impute synthetic data of different nature and the results are discussed. In addition, a case study concerning the prediction (classification) of US market instability (BEAR or BULL) using a data set with mixed missingness mechanisms and mixed nature is presented to evaluate how different types of imputation methods can affect the final results of the classification task.