Working with missing data is the most critical since many ML & DL methods and algorithms do not know how to work with them. It is necessary to restore them before training. Who has the most successful ways to restore them?
Personally, I use the Python library DataWig (https://github.com/awslabs/datawig)
MICE is a very good algorithm for imputing missing data. Here’s a very useful reference for handling missing data
https://stefvanbuuren.name/fimd/
MICE supposes that the missing data are Missing at Random (MAR) i.e. that the probability that a value is missing depends only on observed value and can be predicted using them.
Algorythm , that cat handles missing data :
XGBoost supports missing values by default. In tree algorithms, branch directions for missing values are learned during training. Note that the gblinear booster treats missing values as zeros.
When the missing parameter is specifed, values in the input predictor that is equal to missing will be treated as missing and removed. By default it’s set to NaN.
- Badges
- Science topic
- Similar topics
- Correction