Package org.neo4j.gds.ml.training
Class TrainingStatistics
java.lang.Object
org.neo4j.gds.ml.training.TrainingStatistics
-
Constructor Summary
Constructors -
Method Summary
Modifier and TypeMethodDescriptionvoidaddCandidateStats(ModelCandidateStats statistics) voidaddOuterTrainScore(Metric metric, double score) voidaddTestScore(Metric metric, double score) intdoubledoublegetMainMetric(int trial) getTestScore(Metric metric) getTrainStats(Metric metric) getValidationStats(Metric metric) static org.neo4j.gds.core.utils.mem.MemoryEstimationmemoryEstimationStatsMap(int numberOfMetricsSpecifications, int numberOfModelCandidates) static org.neo4j.gds.core.utils.mem.MemoryEstimationmemoryEstimationStatsMap(int numberOfMetricsSpecifications, int numberOfModelCandidates, int numberOfClasses) toMap()Turns this class into a Cypher map, to be returned in a procedure YIELD field.trainMetricsAvg(int trial) validationMetricsAvg(int trial)
-
Constructor Details
-
TrainingStatistics
-
-
Method Details
-
getTrainStats
-
getValidationStats
-
getTestScore
-
toMap
Turns this class into a Cypher map, to be returned in a procedure YIELD field. This is intentionally omitting the test scores. These can be added to extend the return surface later. -
getMainMetric
public double getMainMetric(int trial) -
validationMetricsAvg
-
trainMetricsAvg
-
evaluationMetric
-
addCandidateStats
-
addTestScore
-
addOuterTrainScore
-
winningModelTestMetrics
-
winningModelOuterTrainMetrics
-
getBestTrialIdx
public int getBestTrialIdx() -
bestCandidate
-
getBestTrialScore
public double getBestTrialScore() -
bestParameters
-
memoryEstimationStatsMap
public static org.neo4j.gds.core.utils.mem.MemoryEstimation memoryEstimationStatsMap(int numberOfMetricsSpecifications, int numberOfModelCandidates) -
memoryEstimationStatsMap
public static org.neo4j.gds.core.utils.mem.MemoryEstimation memoryEstimationStatsMap(int numberOfMetricsSpecifications, int numberOfModelCandidates, int numberOfClasses)
-