Package es.upm.etsisi.cf4j.recommender.neural

Class NeuMF

java.lang.Object

es.upm.etsisi.cf4j.recommender.Recommender

es.upm.etsisi.cf4j.recommender.neural.NeuMF

public class NeuMF
extends Recommender

Implements He, X., Liao, L., Zhang, H., Nie, L., Hu, X., & Chua, T. S. (2017, April). Neural collaborative filtering. In Proceedings of the 26th international conference on world wide web (pp. 173-182).

Field Summary

Fields

Modifier and Type	Field	Description
protected int[]	layers	Array of layers neurons
protected double	learningRate	Learning Rate
protected int	numFactorsGMF	Number of GMF latent factors
protected int	numFactorsMLP	Number of MLP latent factors

Fields inherited from class es.upm.etsisi.cf4j.recommender.Recommender

datamodel

Constructor Summary

Constructors

Constructor	Description
NeuMF(DataModel datamodel, int numEpochs, double learningRate)	Model constructor
NeuMF(DataModel datamodel, int numEpochs, double learningRate, int[] layers)	Model constructor
NeuMF(DataModel datamodel, int numEpochs, double learningRate, long seed)	Model constructor
NeuMF(DataModel datamodel, int numFactorsGMF, int numFactorsMLP, int numEpochs, double learningRate)	Model constructor
NeuMF(DataModel datamodel, int numFactorsGMF, int numFactorsMLP, int numEpochs, double learningRate, int[] layers)	Model constructor
NeuMF(DataModel datamodel, int numFactorsGMF, int numFactorsMLP, int numEpochs, double learningRate, int[] layers, long seed)	Model constructor
NeuMF(DataModel datamodel, int numFactorsGMF, int numFactorsMLP, int numEpochs, double learningRate, long seed)	Model constructor
NeuMF(DataModel datamodel, Map<String,Object> params)	Model constructor from a Map containing the model's hyper-parameters values.

Method Summary

Modifier and Type	Method	Description
void	fit()	Estimates model parameters given the hyper-parameters
int	getGMFNumFactors()	Returns the number of latent factors of GMF net.
int[]	getLayers()	Returns net layers.
double	getLearningRate()	Returns learning rate.
int	getMLPNumFactors()	Returns the number of latent factors of MLP net.
int	getNumEpochs()	Returns the number of epochs.
double	predict(int userIndex, int itemIndex)	Returns the prediction of a rating of a certain user for a certain item, through these predictions the metrics of MAE, MSE and RMSE can be obtained.
String	toString()

Methods inherited from class es.upm.etsisi.cf4j.recommender.Recommender

getDataModel, predict

Methods inherited from class java.lang.Object

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

Field Detail

numFactorsGMF

protected final int numFactorsGMF

Number of GMF latent factors

numFactorsMLP

protected final int numFactorsMLP

Number of MLP latent factors

learningRate

protected final double learningRate

Learning Rate

layers

protected final int[] layers

Array of layers neurons

Constructor Detail

NeuMF

public NeuMF(DataModel datamodel,
             Map<String,Object> params)

Model constructor from a Map containing the model's hyper-parameters values. Map object must contains the following keys:

• numFactorsGMF (optional): int value with the number of latent factors of GMF net. If missing, default 10 GMF latent factors are used.
• numFactorsMLP (optional): int value with the number of latent factors of MLP net. If missing, default 10 MLP latent factors are used.
• numEpochs: int value with the number of epochs.
• learningRate: double value with the learning rate.
• layers (optional): Array of layers neurons. If missing, default [10] Array is used.
• seed (optional): random seed for random numbers generation. If missing, random value is used.

Parameters:

datamodel - DataModel instance

params - Model's hyper-parameters values

NeuMF

public NeuMF(DataModel datamodel,
             int numFactorsGMF,
             int numFactorsMLP,
             int numEpochs,
             double learningRate)

Model constructor

Parameters:

datamodel - DataModel instance

numFactorsGMF - Number of factors of GMF net. 10 by default

numFactorsMLP - Number of factors of MLP net. 10 by default

datamodel - DataModel instance

numEpochs - Number of epochs

learningRate - Learning rate

NeuMF

public NeuMF(DataModel datamodel,
             int numFactorsGMF,
             int numFactorsMLP,
             int numEpochs,
             double learningRate,
             long seed)

Model constructor

Parameters:

datamodel - DataModel instance

numFactorsGMF - Number of factors of GMF net. 10 by default

numFactorsMLP - Number of factors of MLP net. 10 by default

numEpochs - Number of epochs

learningRate - Learning rate

seed - random seed for random numbers generation

NeuMF

public NeuMF(DataModel datamodel,
             int numFactorsGMF,
             int numFactorsMLP,
             int numEpochs,
             double learningRate,
             int[] layers)

Model constructor

Parameters:

datamodel - DataModel instance

numFactorsGMF - Number of factors of GMF net. 10 by default

numFactorsMLP - Number of factors of MLP net. 10 by default

numEpochs - Number of epochs

learningRate - Learning rate

layers - Array of layers neurons. [10] by default.

NeuMF

public NeuMF(DataModel datamodel,
             int numEpochs,
             double learningRate)

Model constructor

Parameters:

datamodel - DataModel instance

datamodel - DataModel instance

numEpochs - Number of epochs

learningRate - Learning rate

NeuMF

public NeuMF(DataModel datamodel,
             int numEpochs,
             double learningRate,
             long seed)

Model constructor

Parameters:

datamodel - DataModel instance

numEpochs - Number of epochs

learningRate - Learning rate

seed - random seed for random numbers generation

NeuMF

public NeuMF(DataModel datamodel,
             int numEpochs,
             double learningRate,
             int[] layers)

Model constructor

Parameters:

datamodel - DataModel instance

numEpochs - Number of epochs

learningRate - Learning rate

layers - Array of layers neurons. [10] by default.

NeuMF

public NeuMF(DataModel datamodel,
             int numFactorsGMF,
             int numFactorsMLP,
             int numEpochs,
             double learningRate,
             int[] layers,
             long seed)

Model constructor

Parameters:

datamodel - DataModel instance

numEpochs - Number of epochs

learningRate - Learning rate

layers - Array of layers neurons. [10] by default.

seed - Seed for random numbers generation

Method Detail

fit

public void fit()

Description copied from class: Recommender

Estimates model parameters given the hyper-parameters

Specified by:

fit in class Recommender

predict

public double predict(int userIndex,
                      int itemIndex)

Returns the prediction of a rating of a certain user for a certain item, through these predictions the metrics of MAE, MSE and RMSE can be obtained.

Specified by:

predict in class Recommender

Parameters:

userIndex - Index of the user in the array of Users of the DataModel instance

itemIndex - Index of the item in the array of Items of the DataModel instance

Returns:

Prediction

getNumEpochs

public int getNumEpochs()

Returns the number of epochs.

Returns:

Number of epochs.

getGMFNumFactors

public int getGMFNumFactors()

Returns the number of latent factors of GMF net.

Returns:

Number of latent factors of GMF net.

getMLPNumFactors

public int getMLPNumFactors()

Returns the number of latent factors of MLP net.

Returns:

Number of latent factors of MLP net.

getLearningRate

public double getLearningRate()

Returns learning rate.

Returns:

learning rate.

getLayers

public int[] getLayers()

Returns net layers.

Returns:

net layers.

toString

public String toString()

Overrides:

toString in class Object