Techniques for Handling Missing Values in Customers Electricity Data: A Systematic Literature Review

The lack of complete data is a widespread challenge when analyzing authentic, real-world datasets. In fact, the missing values can substantially degrade the precision and dependability of statistical examinations, which are foundational in modeling approaches. Effective imputation methods are essential, particularly for time series data, to maintain data integrity. In this paper, various imputation techniques tailored for electricity data are explored, which examines various methods ranging from traditional statistical approaches to advanced deep learning models. The review highlights the advantages and disadvantages of various methods, with particular emphasis on the performance of techniques like Makima Interpolation and LSTM-BIT in handling missing electricity data with a high number of missing values. The study also discusses the common types of missing data and the relevant evaluation measures used to assess imputation performance. In addition, the review aims to provide a clear understanding of the current state of research and practical implications for selecting appropriate imputation methods in electricity data contexts.