hex

Class NeuralNet.NeuralNetModel

java.lang.Object

water.Iced

water.Lockable<Model>

water.Model

hex.NeuralNet.NeuralNetModel

All Implemented Interfaces:

java.lang.Cloneable, Freezable

Enclosing class:

NeuralNet

public static class NeuralNet.NeuralNetModel
extends Model

Nested Class Summary

Nested classes/interfaces inherited from class water.Model

Model.ModelCategory

Field Summary

Fields

Modifier and Type	Field and Description
float[][]	biases
long[][]	confusion_matrix
static DocGen.FieldDoc[]	DOC_FIELDS
Layer[]	layers
float[]	mean_bias
float[]	mean_weight
NeuralNet	parameters
float[]	rms_bias
float[]	rms_weight
NeuralNet.Errors[]	training_errors
boolean	unstable
NeuralNet.Errors[]	validation_errors
float[][]	weights

Fields inherited from class water.Model

_dataKey, _domains, _modelClassDist, _names, _priorClassDist, training_duration_in_ms, training_start_time

Fields inherited from class water.Lockable

_key, _lockers

Method Summary

Methods

Modifier and Type	Method and Description
ConfusionMatrix	cm() For classifiers, confusion matrix on validation set.
boolean	generateHTML(java.lang.String title, java.lang.StringBuilder sb)
protected float[]	score0(Chunk[] chunks, int rowInChunk, double[] tmp, float[] preds) Bulk scoring API for one row.
protected float[]	score0(double[] data, float[] preds) Subclasses implement the scoring logic.
void	toJavaHtml(java.lang.StringBuilder sb)
java.lang.String	toString()

Methods inherited from class water.Model

adapt, calcError, classNames, delete_impl, errStr, get_params, getDomainMapping, getDomainMapping, getModelCategory, getUniqueId, isClassifier, job, missingColumnsType, mse, nclasses, nfeatures, responseName, score, score, score, score, score, score, setModelClassDistribution, start_training, start_training, stop_training, testJavaScoring, toJava, toJava, toJavaDefaultMaxIters, toJavaInit, toJavaInit, toJavaPredictBody, toJavaSuper, varimp

Methods inherited from class water.Lockable

delete_and_lock, delete, delete, delete, delete, is_unlocked, is_wlocked, read_lock, read_lock, unlock_all, unlock, update, write_lock

Methods inherited from class water.Iced

clone, frozenType, init, newInstance, read, toDocField, write, writeJSON, writeJSONFields

Methods inherited from class java.lang.Object

equals, finalize, getClass, hashCode, notify, notifyAll, wait, wait, wait

Field Detail

DOC_FIELDS

public static DocGen.FieldDoc[] DOC_FIELDS

parameters

@Request.API(help="Model parameters")
public NeuralNet parameters

layers

public Layer[] layers

weights

public float[][] weights

biases

public float[][] biases

training_errors

@Request.API(help="Errors on the training set")
public NeuralNet.Errors[] training_errors

validation_errors

@Request.API(help="Errors on the validation set")
public NeuralNet.Errors[] validation_errors

confusion_matrix

@Request.API(help="Confusion matrix")
public long[][] confusion_matrix

mean_bias

@Request.API(help="Mean bias")
public float[] mean_bias

rms_bias

@Request.API(help="RMS bias")
public float[] rms_bias

mean_weight

@Request.API(help="Mean weight")
public float[] mean_weight

rms_weight

@Request.API(help="RMS weight")
public float[] rms_weight

unstable

@Request.API(help="Unstable")
public boolean unstable

Method Detail

toString

public java.lang.String toString()

Overrides:

toString in class java.lang.Object

toJavaHtml

public void toJavaHtml(java.lang.StringBuilder sb)

generateHTML

public boolean generateHTML(java.lang.String title,
                   java.lang.StringBuilder sb)

score0

protected float[] score0(Chunk[] chunks,
             int rowInChunk,
             double[] tmp,
             float[] preds)

Description copied from class: Model

Bulk scoring API for one row. Chunks are all compatible with the model, and expect the last Chunks are for the final distribution and prediction. Default method is to just load the data into the tmp array, then call subclass scoring logic.

Overrides:

score0 in class Model

score0

protected float[] score0(double[] data,
             float[] preds)

Description copied from class: Model

Subclasses implement the scoring logic. The data is pre-loaded into a re-used temp array, in the order the model expects. The predictions are loaded into the re-used temp array, which is also returned.

Specified by:

score0 in class Model

cm

public ConfusionMatrix cm()

Description copied from class: Model

For classifiers, confusion matrix on validation set.

Overrides:

cm in class Model