Icon

Comparing Missing Value Handling Methods

This workflows compares four missing value handling methods
- Listwise Deletion
- 0 imputatition
- Mean & most frequent impuation
- Linear regression & kNN
for two datasets
- Churn Dataset
- Census Dataset

Define how many valuesshould be set to missing inthe different iterations (inpercentage) In each iteration the loop:- Reads the datasets and sets the defined percentage of values to missing. - Imputs the missing values.- Trains, applies, and evaluates a decision tree model. 80% training20% testing Churn datasetCalculate accuracyand Cohen's Kappa Census income dataset80% training20% testingChurn datasetCensus income datasetLoop initializationCalculate accuracyand Cohen's Kappa Partitioning Read data and sprinklemissing values Evaluate models Loop End Table Row ToVariable Loop Start Read data and sprinklemissing values Partitioning Concatenate Impute missing values andtrain and apply models Impute missing values andtrain and apply models Table Creator Evaluate models Visualization Define how many valuesshould be set to missing inthe different iterations (inpercentage) In each iteration the loop:- Reads the datasets and sets the defined percentage of values to missing. - Imputs the missing values.- Trains, applies, and evaluates a decision tree model. 80% training20% testingChurn datasetCalculate accuracyand Cohen's Kappa Census income dataset80% training20% testingChurn datasetCensus income datasetLoop initializationCalculate accuracyand Cohen's Kappa Partitioning Read data and sprinklemissing values Evaluate models Loop End Table Row ToVariable Loop Start Read data and sprinklemissing values Partitioning Concatenate Impute missing values andtrain and apply models Impute missing values andtrain and apply models Table Creator Evaluate models Visualization

Nodes

Extensions

Links