evaluation

Type Members

1. case class ClusteringDistanceSummary(predictions: RDD[(Int, Double)]) extends Product with Serializable

   

   Calculate distance statistics, by cluster and for the entire model.

   Calculate distance statistics, by cluster and for the entire model.

   predictions

   the tuple of cluster id and distance used to compute the statistics

2. case class ClusteringFeaturesSummary(predictions: RDD[(Int, Vector)], centroids: Array[Vector]) extends Product with Serializable

   

   Calculate distance statistics for each feature, by cluster and for the entire model.

   Calculate distance statistics for each feature, by cluster and for the entire model.

   This provides a deeper insight into the model itself than the ClusteringStats

   predictions

   the tuple of cluster id and distance vector used to compute the statistics

3. case class DistanceStatsByFeature(count: Long, min: Vector, avg: Vector, max: Vector, variance: Vector) extends Product with Serializable

   

   Basic statistics for each dimension (minimum, mean, maximum and variance)