A Survey on Missing Values Handling Methods for Time Series Data

A time series is a collection of observations obtained through repeated measurements. The goal of time series analysis is the prediction of future trends based on historical records. The quality of prediction strongly depends on the quality of data. Often a problem is faced with the missing values due to the anomalies during data collection and storage. In this paper, some of the recent methods of data imputation have been discussed that keeps the characteristics of data distribution intact producing high-quality data which can be used more efficiently to arrive at good predictions. The aim of the discussion is to provide a brief knowledge about recent data imputation techniques that will provide researchers a large set of tools to help them choose the right tool to tackle the problem of data missingness.