What is the name of the process used to fill in missing data values using statistical calculations?

The process of filling in missing data values using statistical calculations is known as imputation. Imputation involves using various techniques to estimate and replace the missing data points within a dataset to ensure that the analysis can proceed smoothly without having to discard entire records with incomplete data.

By applying imputation methods, data scientists can maintain the integrity of their datasets and also improve the reliability of their analyses and model predictions. Different types of imputation techniques may involve using the mean, median, or mode of the available data, more complex algorithms like k-nearest neighbors, or even predictive models to fill in those gaps.

While terms like substitution, estimation, and interpolation may seem similar, they do not specifically refer to the standardized statistical process of handling missing values. Substitution often implies replacing a missing value directly with another specific value without statistical backing. Estimation is broader and encompasses various methods for making calculated guesses about unknown values but does not specifically refer to missing data. Interpolation, on the other hand, is a technique used to estimate unknown values between two known values in a dataset but is not used primarily to handle missing data points. Hence, the use of the term imputation is precise in the context of addressing missing values in data.