Missing Values imputation component - ClickML Studio

The Missing Values component extends basic cleaning with more imputation strategies, including KNN-based imputation and interpolation.

Configuration

Option	Description
Columns	Columns to apply the strategy to. Supports `All Numerical Features` and `All Categorical Features` shortcuts. Leave empty to apply to all columns.
Strategy	Imputation method (see table below).
Fill Value	The constant value to use when Strategy is `Fill Constant`.
N Neighbors	Number of neighbors for `KNN Imputation` (default: 5).

Strategies

Strategy	Description
Drop Rows	Remove rows with any missing value in selected columns
Drop Columns	Remove the selected columns
Fill Constant	Fill missing values with a fixed value
Fill Mean	Fill with column mean
Fill Median	Fill with column median
Fill Mode	Fill with most frequent value
Fill Forward	Propagate the last valid value forward
Fill Backward	Propagate the next valid value backward
Interpolate Linear	Linear interpolation between known values (numerical only)
KNN Imputation	Impute using K-nearest neighbors based on other features (numerical only)

Input / Output

	Type
Input	DataFrame
Output	DataFrame

KNN Imputation produces the most accurate results but is slower on large datasets. For quick fixes, Fill Mean or Fill Median is usually sufficient.