H2O Clustering Evaluator

Evaluator for Clustering, which expects two input columns: features and prediction.

Output

It outputs the silhouette score

Type

ml-evaluator

Class

fire.nodes.h2o.NodeH2OClusteringEvaluator

Fields

Name

Title

Description

predictionCol

Prediction Column

The prediction column.

metric

Metric Column

The distance metric to use for evaluating the clustering algorithm. This can be set to ‘cosine’, or ‘squaredEuclidean

Details

Node H2O Clustering Evaluator Details:

Evaluator for Clustering , which expects three input columns: prediction,metric and label.

Input Parameters

  • OUTPUT STORAGE LEVEL : Keep this as DEFAULT.

  • Prediction Column : The prediction column created during model scoringThe name of the column containing predictions generated by the clustering algorithm..

  • Metric Column : The distance metric to use for evaluating the clustering algorithm. This can be set to ‘cosine’, ‘squaredEuclidean’.