Data Imputation

Data imputation simply means replacing the missing data with something.

There are multiple ways to impute data,

  1. Mean
  2. Median
  3. Mode
  4. Replace with most co-related data (see Finding Co-relation between two data or distribution)
  5. Assign new category i.e. unknown (For categorical values)
  6. Interpolation (Time series data)
  7. Use K-nearest Neighbor (KNN) to interpolate data

Related Notes