SimpleImputer<T>

Models & Types · AiDotNet.Preprocessing.Imputers

Imputes missing values using simple strategies like mean, median, or constant.

Inheritance: TransformerBase<T, Matrix<T>, Matrix<T>> → SimpleImputer<T>

For Beginners

This transformer fills in gaps in your data:

  • If you have missing ages, replace them with average age (Mean)
  • If you have missing incomes with outliers, use median income (Median)
  • If you have missing categories, use most common category (MostFrequent)
  • Or fill with a specific value like 0 or -1 (Constant)

Example with Mean strategy: [1, 2, NaN, 4, 5] → [1, 2, 3, 4, 5] (NaN replaced with mean=3)

How It Works

SimpleImputer fills in missing values (represented as NaN) using simple strategies:

  • Mean: Replace with column mean
  • Median: Replace with column median
  • MostFrequent: Replace with most common value
  • Constant: Replace with a specified value

Constructors

Constructor Summary
SimpleImputer(ImputationStrategy strategy, T fillValue, int[] columnIndices) Creates a new instance of SimpleImputer.

Properties

Property Summary
ImputationStrategy Strategy { get; } Gets the imputation strategy used.
T FillValue { get; } Gets the fill value used for Constant strategy.
T MissingValue { get; } Gets the value considered as missing.
Vector<T> Statistics { get; } Gets the computed statistics for each feature.
bool SupportsInverseTransform { get; } Gets whether this transformer supports inverse transformation.

Methods

Method Summary
string[] GetFeatureNamesOut(string[] inputFeatureNames) Gets the output feature names after transformation.