Package org.deeplearning4j.eval

Interface Summary

Interface	Description
IEvaluation<T extends IEvaluation>	A general purpose interface for evaluating neural networks - methods are shared by implemetations such as Evaluation, RegressionEvaluation, ROC, ROCMultiClass

Class Summary

Class	Description
BaseEvaluation<T extends BaseEvaluation>	BaseEvaluation implement common evaluation functionality (for time series, etc) for Evaluation, RegressionEvaluation, ROC, ROCMultiClass etc.
ConfusionMatrix<T extends Comparable<? super T>>
Evaluation	Evaluation metrics: - precision, recall, f1, fBeta, accuracy, Matthews correlation coefficient, gMeasure - Top N accuracy (if using constructor Evaluation.Evaluation(List, int)) - Custom binary evaluation decision threshold (use constructor Evaluation.Evaluation(double) (default if not set is argmax / 0.5) - Custom cost array, using Evaluation.Evaluation(INDArray) or Evaluation.Evaluation(List, INDArray) for multi-class Note: Care should be taken when using the Evaluation class for binary classification metrics such as F1, precision, recall, etc.
EvaluationBinary	EvaluationBinary: used for evaluating networks with binary classification outputs.
EvaluationCalibration	EvaluationCalibration is an evaluation class designed to analyze the calibration of a classifier. It provides a number of tools for this purpose: - Counts of the number of labels and predictions for each class - Reliability diagram (or reliability curve) - Residual plot (histogram) - Histograms of probabilities, including probabilities for each class separately References: - Reliability diagram: see for example Niculescu-Mizil and Caruana 2005, Predicting Good Probabilities With Supervised Learning - Residual plot: see Wallace and Dahabreh 2012, Class Probability Estimates are Unreliable for Imbalanced Data (and How to Fix Them)
EvaluationUtils	Utility methods for performing evaluation
RegressionEvaluation	Evaluation method for the evaluation of regression algorithms. Provides the following metrics, for each column: - MSE: mean squared error - MAE: mean absolute error - RMSE: root mean squared error - RSE: relative squared error - PC: pearson correlation coefficient - R^2: coefficient of determination See for example: http://www.saedsayad.com/model_evaluation_r.htm For classification, see Evaluation
ROC	ROC (Receiver Operating Characteristic) for binary classifiers. ROC has 2 modes of operation: (a) Thresholded (less memory) (b) Exact (default; use numSteps == 0 to set.
ROC.CountsForThreshold
ROCBinary	ROC (Receiver Operating Characteristic) for multi-task binary classifiers.
ROCMultiClass	ROC (Receiver Operating Characteristic) for multi-class classifiers.

Enum Summary

Enum	Description
Evaluation.Metric
EvaluationAveraging	The averaging approach for binary valuation measures when applied to multiclass classification problems.
RegressionEvaluation.Metric