Advancing Missing Data Imputation in Time-Series: A Review and Proposed Prototype

Completeness and quality of data are crucial for any type of data analysis and play a vital role in time series forecasting. However, a common challenge in time series forecasting is dealing with missing values, which can harm the accuracy and reliability of forecasting tasks. If not handled properly, these missing values can disrupt the continuity of the time series and lead to inaccurate predictions. Therefore, addressing the issue of missing values is essential to ensure the effectiveness and reliability of time series forecasting. Thus, the primary scope of this research is to explore different methods for missing data imputation in time-series data. The paper also addresses open challenges and opportunities for future research, including tackling specific types of missingness patterns, handling high-dimensional time-stamped data, and incorporating domain knowledge into the imputation process. In the end, a prototype is proposed that outlines the essential steps involved in constructing a model for the imputation of missing values in time-stamped data. A thorough discussion is presented on the implications of the findings and the potential avenues for future research in the context of time-stamped data. Finally, the key findings and future directions for advancing the field of missing value imputation in time-stamped data are also discussed.