Handles missing data

Option	Description
Yes	Built-in handling of missing values (e.g., likelihood‐based, imputation).
Partially	Limited capacity to handle missing data, e.g. by automatically discarding incomplete time series.
No: requires prelim. correction	Cannot function without complete data and therefore requires the data to be imputed before applying.

Definition

Whether the method can accommodate or adjust for missing values in the dataset.

Explanation

Missing data are common in RS and ecological datasets; methods must either handle gaps or require imputation.

Tools/rationale for helping assessment

Count the number of incomplete variables in the system relative to the total variable number, and the percentage of missing values per such variable. If the user estimates unreasonable to drop these variables, the next step is to specify how this aspect should be taken into account by the method: It is central, suggested methods should have built-in capacities (Yes), It is secondary, i.e. simple tools to leverage the data are enough (Partially), and finally externally: imputation of missing data should be done with a preliminary algorithm.

Example

Cloud cover masks ~30 % of my Landsat images. I need a method for handling such large data gaps with built-in capacities, so I select Yes (e.g. state‐space gap‐filling).