Time series signal recovery methods: comparative study

TL;DR

This study compares three common methods for imputing missing values in time series data, analyzing their effectiveness across various correlation scenarios through extensive simulations.

Contribution

It provides a comprehensive experimental comparison of forward fill, backward fill, and mean fill methods for time series imputation, clarifying their suitability based on correlation characteristics.

Findings

Forward and backward fill perform better with highly positively correlated series.

Mean fill is more effective for series with low or negative correlations.

Extensive simulations offer a definitive comparison of the three methods.

Abstract

Signal data often contains missing values. Effective replacement (imputation) of the missing values can have significant positive effects on processing the signal. In this paper, we compare three commonly employed methods for estimating missing values in time series data: forward fill, backward fill, and mean fill. We carry out a large scale experimental analysis using 3,600 AR(1)-based simulated time series to determine the optimal method for estimating missing values. The results of the numerical experiments show that the forward and backward fill methods are better suited for times series with large positive correlations, while the mean fill method is better suited for times series with low or negative correlations. The extensive and exhaustive nature of the numerical experiments provides a definitive answer to the comparison of the three imputation methods.