/**
* Copyright 2010 Neuroph Project http://neuroph.sourceforge.net
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
package org.neuroph.contrib;
import org.neuroph.core.Layer;
import org.neuroph.core.NeuralNetwork;
import org.neuroph.core.Neuron;
import org.neuroph.nnet.learning.UnsupervisedHebbianLearning;
import org.neuroph.util.LayerFactory;
import org.neuroph.util.NeuralNetworkFactory;
import org.neuroph.util.NeuralNetworkType;
import org.neuroph.util.NeuronProperties;
import org.neuroph.util.TransferFunctionType;
import org.neuroph.util.plugins.LabelsPlugin;
/**
* E-commerce recommender neural network based on hebbian learning.
* Still under development, mostly hard-coded for now
* @author Zoran Sevarac <sevarac@gmail.com>
*/
public class RecommenderNetwork extends NeuralNetwork {
private static final long serialVersionUID = 1L;
int inputLayerIdx = 0,
typeLayerIdx = 1,
brandLayerIdx = 2,
priceLayerIdx = 3,
promoLayerIdx = 4,
outputLayerIdx = 5;
public RecommenderNetwork() {
for(int i = 0; i < 6; i++)
this.addLayer(new Layer());
//this.createDemoNetwork();
}
public void createDemoNetwork() {
int productsCount = 20;
int typesCount = 3;
int brandsCount = 3;
int priceCount = 3;
int promoCount = 3;
this.setNetworkType(NeuralNetworkType.RECOMMENDER);
this.getLayers().clear();
// init neuron settings for this type of network
NeuronProperties neuronProperties = new NeuronProperties();
neuronProperties.setProperty("transferFunction", TransferFunctionType.RAMP);
// for sigmoid and tanh transfer functions
neuronProperties.setProperty("transferFunction.slope", new Double(1));
// create input layer
Layer inputLayer = LayerFactory.createLayer(productsCount,
neuronProperties);
this.addLayer(inputLayer);
createProductLabels(inputLayer);
// create product types layer
Layer typeLayer = LayerFactory.createLayer(typesCount,
neuronProperties);
createTypeLabels(typeLayer);
this.addLayer(typeLayer);
// create brands layer
Layer brandLayer = LayerFactory.createLayer(brandsCount,
neuronProperties);
createBrandLabels(brandLayer);
this.addLayer(brandLayer);
// create price layer
Layer priceLayer = LayerFactory.createLayer(priceCount, neuronProperties);
createPriceLabels(priceLayer);
this.addLayer(priceLayer);
// create price layer
Layer promoLayer = LayerFactory.createLayer(promoCount, neuronProperties);
createPromoLabels(promoLayer);
this.addLayer(promoLayer);
// create output layer
Layer outputLayer = LayerFactory.createLayer(productsCount,
neuronProperties);
this.addLayer(outputLayer);
createProductLabels(outputLayer);
createTypeConnections();
createBrandConnections();
createPriceConnections();
createPromoConnections();
// create reccurent self connections in output layer
for(Neuron neuron : this.getLayerAt(outputLayerIdx).getNeurons()) {
neuron.addInputConnection(neuron, 1);
}
// set input and output cells for this network
NeuralNetworkFactory.setDefaultIO(this);
// dont learn the self connections
// moze cak i posle svakog prolaza da se primenjuje hebbianovo pravilo a ne samo nakon kupovine
// napravi vise varijanti
// ako kupuje onda moze da se primenjje winner takes all hebbian learning
this.setLearningRule(new UnsupervisedHebbianLearning());
}
private void createTypeConnections() {
// input connections for the first product type
for(int i = 0; i < 7; i++) {
Neuron fromNeuron = this.getLayerAt(inputLayerIdx).getNeuronAt(i);
Neuron toNeuron = this.getLayerAt(typeLayerIdx).getNeuronAt(0);
this.createConnection(fromNeuron, toNeuron, 0.1);
}
// input connections for the second product type
for(int i = 7; i < 14; i++) {
Neuron fromNeuron = this.getLayerAt(inputLayerIdx).getNeuronAt(i);
Neuron toNeuron = this.getLayerAt(typeLayerIdx).getNeuronAt(1);
this.createConnection(fromNeuron, toNeuron, 0.1);
}
// input connections for the third product type
for(int i = 14; i < 20; i++) {
Neuron fromNeuron = this.getLayerAt(inputLayerIdx).getNeuronAt(i);
Neuron toNeuron = this.getLayerAt(typeLayerIdx).getNeuronAt(2);
this.createConnection(fromNeuron, toNeuron, 0.1);
}
// output connections for the first product type
for(int i = 0; i < 7; i++) {
Neuron fromNeuron = this.getLayerAt(typeLayerIdx).getNeuronAt(0);
Neuron toNeuron = this.getLayerAt(outputLayerIdx).getNeuronAt(i);
this.createConnection(fromNeuron, toNeuron, 0.1);
}
// output connections for the second product type
for(int i = 7; i < 14; i++) {
Neuron fromNeuron = this.getLayerAt(typeLayerIdx).getNeuronAt(1);
Neuron toNeuron = this.getLayerAt(outputLayerIdx).getNeuronAt(i);
this.createConnection(fromNeuron, toNeuron, 0.1);
}
// output connections for the third product type
for(int i = 14; i < 20; i++) {
Neuron fromNeuron = this.getLayerAt(typeLayerIdx).getNeuronAt(2);
Neuron toNeuron = this.getLayerAt(outputLayerIdx).getNeuronAt(i);
this.createConnection(fromNeuron, toNeuron, 0.1);
}
}
private void createBrandConnections() {
int samsung[] = {0, 1, 7, 8, 9, 14, 15};
int lg[] = {2, 3, 10, 11, 16, 17};
int sony[] = {4, 5, 6, 12, 13, 18, 19};
// create samsung input sonnections
for(int i = 0; i < 7; i++) {
Neuron fromNeuron = this.getLayerAt(inputLayerIdx).getNeuronAt(samsung[i]);
Neuron toNeuron = this.getLayerAt(brandLayerIdx).getNeuronAt(0);
this.createConnection(fromNeuron, toNeuron, 0.1);
}
// create input connections for LG
for(int i = 0; i < 6; i++) {
Neuron fromNeuron = this.getLayerAt(inputLayerIdx).getNeuronAt(lg[i]);
Neuron toNeuron = this.getLayerAt(brandLayerIdx).getNeuronAt(1);
this.createConnection(fromNeuron, toNeuron, 0.1);
}
// create input connections for sony
for(int i = 0; i < 7; i++) {
Neuron fromNeuron = this.getLayerAt(inputLayerIdx).getNeuronAt(sony[i]);
Neuron toNeuron = this.getLayerAt(brandLayerIdx).getNeuronAt(2);
this.createConnection(fromNeuron, toNeuron, 0.1);
}
// output connections for the first brand
for(int i = 0; i < 7; i++) {
Neuron fromNeuron = this.getLayerAt(brandLayerIdx).getNeuronAt(0);
Neuron toNeuron = this.getLayerAt(outputLayerIdx).getNeuronAt(samsung[i]);
this.createConnection(fromNeuron, toNeuron, 0.1);
}
// output connections for the second brand
for(int i = 0; i < 6; i++) {
Neuron fromNeuron = this.getLayerAt(brandLayerIdx).getNeuronAt(1);
Neuron toNeuron = this.getLayerAt(outputLayerIdx).getNeuronAt(lg[i]);
this.createConnection(fromNeuron, toNeuron, 0.1);
}
// output connections for the third brand
for(int i = 0; i < 7; i++) {
Neuron fromNeuron = this.getLayerAt(brandLayerIdx).getNeuronAt(2);
Neuron toNeuron = this.getLayerAt(outputLayerIdx).getNeuronAt(sony[i]);
this.createConnection(fromNeuron, toNeuron, 0.1);
}
}
private void createPriceConnections() {
int low[] = {0, 2, 4, 7, 10, 16, 18};
int mid[] = {3, 5, 8, 11, 12, 14, 19};
int high[] = {1, 6, 9, 13, 15, 17};
// input connections for the first price class
for(int i = 0; i < 7; i++) {
Neuron fromNeuron = this.getLayerAt(inputLayerIdx).getNeuronAt(low[i]);
Neuron toNeuron = this.getLayerAt(priceLayerIdx).getNeuronAt(0);
this.createConnection(fromNeuron, toNeuron, 0.1);
}
// input connections for the second price class
for(int i = 0; i < 7; i++) {
Neuron fromNeuron = this.getLayerAt(inputLayerIdx).getNeuronAt(mid[i]);
Neuron toNeuron = this.getLayerAt(priceLayerIdx).getNeuronAt(1);
this.createConnection(fromNeuron, toNeuron, 0.1);
}
// input connections for the third price class
for(int i = 0; i < 6; i++) {
Neuron fromNeuron = this.getLayerAt(inputLayerIdx).getNeuronAt(high[i]);
Neuron toNeuron = this.getLayerAt(priceLayerIdx).getNeuronAt(2);
this.createConnection(fromNeuron, toNeuron, 0.1);
}
// output connections for the first price class
for(int i = 0; i < 7; i++) {
Neuron fromNeuron = this.getLayerAt(priceLayerIdx).getNeuronAt(0);
Neuron toNeuron = this.getLayerAt(outputLayerIdx).getNeuronAt(low[i]);
this.createConnection(fromNeuron, toNeuron, 0.1);
}
// output connections for the second price class
for(int i = 0; i < 7; i++) {
Neuron fromNeuron = this.getLayerAt(priceLayerIdx).getNeuronAt(1);
Neuron toNeuron = this.getLayerAt(outputLayerIdx).getNeuronAt(mid[i]);
this.createConnection(fromNeuron, toNeuron, 0.1);
}
// output connections for the third price class
for(int i = 0; i < 6; i++) {
Neuron fromNeuron = this.getLayerAt(priceLayerIdx).getNeuronAt(2);
Neuron toNeuron = this.getLayerAt(outputLayerIdx).getNeuronAt(high[i]);
this.createConnection(fromNeuron, toNeuron, 0.1);
}
}
private void createPromoConnections() {
int sales[] = {0, 10, 19};
int new_products[] = {6, 9};
int bestsellers[] = {3, 12, 14};
// input connections for the first promo type
for(int i = 0; i < sales.length; i++) {
Neuron fromNeuron = this.getLayerAt(inputLayerIdx).getNeuronAt(sales[i]);
Neuron toNeuron = this.getLayerAt(promoLayerIdx).getNeuronAt(0);
this.createConnection(fromNeuron, toNeuron, 0.1);
}
// input connections for the second promo type
for(int i = 0; i < new_products.length; i++) {
Neuron fromNeuron = this.getLayerAt(inputLayerIdx).getNeuronAt(new_products[i]);
Neuron toNeuron = this.getLayerAt(promoLayerIdx).getNeuronAt(1);
this.createConnection(fromNeuron, toNeuron, 0.1);
}
// input connections for the third promo type
for(int i = 0; i < bestsellers.length; i++) {
Neuron fromNeuron = this.getLayerAt(inputLayerIdx).getNeuronAt(bestsellers[i]);
Neuron toNeuron = this.getLayerAt(promoLayerIdx).getNeuronAt(2);
this.createConnection(fromNeuron, toNeuron, 0.1);
}
// output connections for the first promo type
for(int i = 0; i < sales.length; i++) {
Neuron fromNeuron = this.getLayerAt(promoLayerIdx).getNeuronAt(0);
Neuron toNeuron = this.getLayerAt(outputLayerIdx).getNeuronAt(sales[i]);
this.createConnection(fromNeuron, toNeuron, 0.1);
}
// output connections for the second promo type
for(int i = 0; i < new_products.length; i++) {
Neuron fromNeuron = this.getLayerAt(promoLayerIdx).getNeuronAt(1);
Neuron toNeuron = this.getLayerAt(outputLayerIdx).getNeuronAt(new_products[i]);
this.createConnection(fromNeuron, toNeuron, 0.1);
}
// output connections for the third promo type
for(int i = 0; i < bestsellers.length; i++) {
Neuron fromNeuron = this.getLayerAt(promoLayerIdx).getNeuronAt(2);
Neuron toNeuron = this.getLayerAt(outputLayerIdx).getNeuronAt(bestsellers[i]);
this.createConnection(fromNeuron, toNeuron, 0.1);
}
}
private void createProductLabels(Layer layer) {
LabelsPlugin labelsPlugin = (LabelsPlugin)getPlugin("LabelsPlugin");
labelsPlugin.setLabel(layer.getNeuronAt(0), "Samsung LCD TV LE-32A330");
labelsPlugin.setLabel(layer.getNeuronAt(1), "Samsung LCD TV LE-32A558");
labelsPlugin.setLabel(layer.getNeuronAt(2), "LG LCD TV 32LG2000");
labelsPlugin.setLabel(layer.getNeuronAt(3), "LG LCD TV 32LG5010");
labelsPlugin.setLabel(layer.getNeuronAt(4), "Sony LCD TV KDL-32L4000K");
labelsPlugin.setLabel(layer.getNeuronAt(5), "Sony LCD TV KDL-32S4000");
labelsPlugin.setLabel(layer.getNeuronAt(6), "Sony LCD TV KDL-32W4000K");
labelsPlugin.setLabel(layer.getNeuronAt(7), "Samsung Digital Camera S760");
labelsPlugin.setLabel(layer.getNeuronAt(8), "Samsung Digital Camera L100");
labelsPlugin.setLabel(layer.getNeuronAt(9), "Samsung Digital Camera S850");
labelsPlugin.setLabel(layer.getNeuronAt(10), "LG Digital Camera DMCLS80E");
labelsPlugin.setLabel(layer.getNeuronAt(11), "LG Digital Camera DMCLZ8E");
labelsPlugin.setLabel(layer.getNeuronAt(12), "Sony Digital Camera DSCW120S");
labelsPlugin.setLabel(layer.getNeuronAt(13), "Sony Digital Camera DSCW130S");
labelsPlugin.setLabel(layer.getNeuronAt(14), "Samsung Mobile Phone E251");
labelsPlugin.setLabel(layer.getNeuronAt(15), "Samsung Mobile Phone U600");
labelsPlugin.setLabel(layer.getNeuronAt(16), "Sony Mobile Phone KP100");
labelsPlugin.setLabel(layer.getNeuronAt(17), "Sony Mobile Phone KE850");
labelsPlugin.setLabel(layer.getNeuronAt(18), "LG Mobile Phone K330");
labelsPlugin.setLabel(layer.getNeuronAt(19), "LG Mobile Phone K660");
}
private void createTypeLabels(Layer layer) {
LabelsPlugin labelsPlugin = (LabelsPlugin)getPlugin("LabelsPlugin");
labelsPlugin.setLabel(layer.getNeuronAt(0), "LCD TV");
labelsPlugin.setLabel(layer.getNeuronAt(1), "Digital Camera");
labelsPlugin.setLabel(layer.getNeuronAt(2), "Mobile Phone");
}
private void createBrandLabels(Layer layer) {
LabelsPlugin labelsPlugin = (LabelsPlugin)getPlugin("LabelsPlugin");
labelsPlugin.setLabel(layer.getNeuronAt(0), "Samsung");
labelsPlugin.setLabel(layer.getNeuronAt(1), "LG");
labelsPlugin.setLabel(layer.getNeuronAt(2), "Sony");
}
private void createPriceLabels(Layer layer) {
LabelsPlugin labelsPlugin = (LabelsPlugin)getPlugin("LabelsPlugin");
labelsPlugin.setLabel(layer.getNeuronAt(0), "Low Price");
labelsPlugin.setLabel(layer.getNeuronAt(1), "Mid Price");
labelsPlugin.setLabel(layer.getNeuronAt(2), "High Price");
}
private void createPromoLabels(Layer layer) {
LabelsPlugin labelsPlugin = (LabelsPlugin)getPlugin("LabelsPlugin");
labelsPlugin.setLabel(layer.getNeuronAt(0), "Sales");
labelsPlugin.setLabel(layer.getNeuronAt(1), "New");
labelsPlugin.setLabel(layer.getNeuronAt(2), "Bestseller");
}
}