When examining features in Decanter AI, for each feature you can select the missing value handling methods to be applied. Under [Missing Handling], click on "Auto" to view the drop-down menu and select the imputation method desired for the feature.




For categorical features, the missing value handling options are:


1. Auto: Decanter AI will determine the best imputation technique for this feature.

2. Mode: Impute the most frequent value in this feature

3. N/A: Use N/A for the missing values as a new category

4. Drop Rows: Drop the rows with missing values entirely and not consider it in the training process


For numerical features, the missing value handling  options are:


1. Auto: Decanter AI will determine the best imputation technique for this feature.

2. Mean: Decanter AI will calculate the mean of the non-missing values and replace the missing values with the feature mean.

3. Median: Decanter AI will calculate the median of the non-missing values and replace the missing values with the feature median.

4. 0: Replace missing values with 0

5: Drop Rows: Drop the rows with missing values entirely and not consider it in the training process