Distinction

Load the “Heart_Failure_Clinical_Records.csv” dataset

Convert numerical attributes to categorical attributes.

Suite 1: 50% train / 50% test

Z-score normalization

Choose "Select best score"

Modal J

Connect the first output port of the “Normalizer” node to thefirst input port of the Reference Column Filter node.

Choose "Accuracy"

Include all features except the DEATH_EVENT.Feature selection strategy: RandomUse lower bound for number of features: 2Use upper bound for number of features: 12

Modal K

- “Stratified sampling” with Suite 3: 90% train / 10% testo Set the random seed to 9698.

Modal J's Result

For Draw Randomly Sampling with Suite 3: 90% train / 10% test

Relative (%): 80o Partitioning method: Stratified samplingo Use random seed: 9698

Modal K

Split Criterion: Gini Indexo Number of models: 200o Use static random seed: 9698

Modal J