[marf-cvs] marf/src/marf/Classification/NeuralNetwork NeuralNetwork.java,1.5,1.6 Neuron.java,1.1,1.2
Brought to you by:
mokhov
Menu
▾
▴
[marf-cvs] marf/src/marf/Classification/NeuralNetwork NeuralNetwork.java,1.5,1.6 Neuron.java,1.1,1.2
|
From: <mo...@us...> - 2002-11-23 16:46:59
|
Update of /cvsroot/marf/marf/src/marf/Classification/NeuralNetwork
In directory sc8-pr-cvs1:/tmp/cvs-serv9540/NeuralNetwork
Modified Files:
NeuralNetwork.java Neuron.java
Log Message:
Update of frameworked NN. Ian, if you're
working on it, incorporate some chages in your
working copy.
Index: NeuralNetwork.java
===================================================================
RCS file: /cvsroot/marf/marf/src/marf/Classification/NeuralNetwork/NeuralNetwork.java,v
retrieving revision 1.5
retrieving revision 1.6
diff -C2 -d -r1.5 -r1.6
*** NeuralNetwork.java 19 Nov 2002 13:04:11 -0000 1.5
--- NeuralNetwork.java 23 Nov 2002 16:46:56 -0000 1.6
***************
*** 17,24 ****
/**
* Class NeuralNetwork
*
! * $Header$
*/
-
public class NeuralNetwork extends Classification
{
--- 17,24 ----
/**
* Class NeuralNetwork
+ * @author Ian Clement
*
! * <p>$Header$
*/
public class NeuralNetwork extends Classification
{
***************
*** 30,38 ****
private int currLayerBuf=0;
private Neuron currNeuron;
! private int neuron_type = -1;
private ArrayList inputs = new ArrayList();
private ArrayList outputs = new ArrayList();
!
//All output will use this encoding
static final String outputEncoding = "UTF-8";
--- 30,38 ----
private int currLayerBuf=0;
private Neuron currNeuron;
! private int neuron_type = -1;
private ArrayList inputs = new ArrayList();
private ArrayList outputs = new ArrayList();
!
//All output will use this encoding
static final String outputEncoding = "UTF-8";
***************
*** 47,88 ****
! public NeuralNetwork(FeatureExtraction poFeatureExtraction)
! {
! super(poFeatureExtraction);
! }
public boolean classify() throws ClassificationException
{
! double [] adFeatures = oFeatureExtraction.getFeatures();
! if(adFeatures.length != inputs.size())
! throw new ClassificationException("Input array size not consistent with input layer.");
!
! for(int i=0; i<adFeatures.length; i++)
! ((Neuron)inputs.get(i)).result = adFeatures[i];
!
! ArrayList current = null;
!
! for(int i=0; i<layers.size(); i++) {
! current = (ArrayList)layers.get(i);
!
! for(int j=0; j<current.size(); j++)
! ((Neuron)current.get(j)).eval();
! }
! //Make result...
! double [] ret = new double[outputs.size()];
! for(int i=0; i<outputs.size(); i++)
! ret[i] =((Neuron)outputs.get(i)).result;
!
! //... and then something to convert this into a result.
! //oResult = (Result)ret;
! }
!
! public Result getResult()
! {
! return oResult;
}
--- 47,86 ----
! public NeuralNetwork(FeatureExtraction poFeatureExtraction)
! {
! super(poFeatureExtraction);
! }
public boolean classify() throws ClassificationException
{
! double [] adFeatures = oFeatureExtraction.getFeatures();
! if(adFeatures.length != inputs.size())
! throw new ClassificationException("Input array size not consistent with input layer.");
! for(int i=0; i<adFeatures.length; i++)
! ((Neuron)inputs.get(i)).result = adFeatures[i];
! ArrayList current = null;
! for(int i=0; i<layers.size(); i++) {
! current = (ArrayList)layers.get(i);
!
! for(int j=0; j<current.size(); j++)
! ((Neuron)current.get(j)).eval();
! }
!
! //Make result...
! double [] ret = new double[outputs.size()];
!
! for(int i=0; i<outputs.size(); i++)
! ret[i] =((Neuron)outputs.get(i)).result;
!
! //... and then something to convert this into a result.
! //oResult = (Result)ret;
! // TODO:
! oResult = new Result(ret);
!
! return (outputs.size() > 0);
}
***************
*** 144,151 ****
createLinks(doc);
}
!
! //DOM tree traversal -- build NNet structure
private void buildNNet(Node n) {
!
int type = n.getNodeType();
String name, value;
--- 142,149 ----
createLinks(doc);
}
!
! //DOM tree traversal -- build NNet structure
private void buildNNet(Node n) {
!
int type = n.getNodeType();
String name, value;
***************
*** 166,170 ****
for (int i = 0; i < atts.getLength(); i++) {
Node att = atts.item(i);
!
String attName = att.getNodeName();
String attValue = att.getNodeValue();
--- 164,168 ----
for (int i = 0; i < atts.getLength(); i++) {
Node att = atts.item(i);
!
String attName = att.getNodeName();
String attValue = att.getNodeValue();
***************
*** 177,181 ****
else if(attValue.equals("output")) {
curr = outputs;
! neuron_type = 2;
}
else {
--- 175,179 ----
else if(attValue.equals("output")) {
curr = outputs;
! neuron_type = 2;
}
else {
***************
*** 194,198 ****
}
else if(name.equals("neuron")) {
!
String neuron_name = new String();
double thresh = 0.0;
--- 192,196 ----
}
else if(name.equals("neuron")) {
!
String neuron_name = new String();
double thresh = 0.0;
***************
*** 200,213 ****
for (int i = 0; i < atts.getLength(); i++) {
Node att = atts.item(i);
!
String attName = att.getNodeName();
String attValue = att.getNodeValue();
!
if(attName.equals("index")) {
System.out.println("Setting neuron name to " + attValue);
neuron_name = new String(attValue);
}
! else if(attName.equals("thresh")) {
! try {
thresh = Double.valueOf(attValue.trim()).doubleValue();
System.out.println("Setting threshold to " + thresh + ".");
--- 198,211 ----
for (int i = 0; i < atts.getLength(); i++) {
Node att = atts.item(i);
!
String attName = att.getNodeName();
String attValue = att.getNodeValue();
!
if(attName.equals("index")) {
System.out.println("Setting neuron name to " + attValue);
neuron_name = new String(attValue);
}
! else if(attName.equals("thresh")) {
! try {
thresh = Double.valueOf(attValue.trim()).doubleValue();
System.out.println("Setting threshold to " + thresh + ".");
***************
*** 222,233 ****
//System.out.println("Making new neuron " + neuron_name + " of type " + type);
!
! Neuron tmp = new Neuron(neuron_name, neuron_type);
tmp.threshold = thresh;
curr.add(tmp);
!
}
}
!
// Recurse for children if any
for (Node child = n.getFirstChild(); child != null; child = child.getNextSibling()) {
--- 220,231 ----
//System.out.println("Making new neuron " + neuron_name + " of type " + type);
!
! Neuron tmp = new Neuron(neuron_name, neuron_type);
tmp.threshold = thresh;
curr.add(tmp);
!
}
}
!
// Recurse for children if any
for (Node child = n.getFirstChild(); child != null; child = child.getNextSibling()) {
***************
*** 235,239 ****
}
}
!
//DOM tree traversal -- create input and output links
private void createLinks(Node n) throws ClassificationException
--- 233,237 ----
}
}
!
//DOM tree traversal -- create input and output links
private void createLinks(Node n) throws ClassificationException
***************
*** 241,258 ****
int type = n.getNodeType();
String name, value;
!
if(type == Node.ELEMENT_NODE) {
!
name = n.getNodeName();
!
NamedNodeMap atts = n.getAttributes();
!
if(name.equals("layer")) {
for (int i = 0; i < atts.getLength(); i++) {
Node att = atts.item(i);
!
String attName = att.getNodeName();
String attValue = att.getNodeValue();
!
if(attName.equals("type")) {
if(attValue.equals("input")) {
--- 239,256 ----
int type = n.getNodeType();
String name, value;
!
if(type == Node.ELEMENT_NODE) {
!
name = n.getNodeName();
!
NamedNodeMap atts = n.getAttributes();
!
if(name.equals("layer")) {
for (int i = 0; i < atts.getLength(); i++) {
Node att = atts.item(i);
!
String attName = att.getNodeName();
String attValue = att.getNodeValue();
!
if(attName.equals("type")) {
if(attValue.equals("input")) {
***************
*** 273,309 ****
}
else if(name.equals("neuron")) {
!
String index = new String();
!
for (int i = 0; i < atts.getLength(); i++) {
Node att = atts.item(i);
!
String attName = att.getNodeName();
String attValue = att.getNodeValue();
!
if(attName.equals("index")) {
index = new String(attValue);
}
}
!
currNeuron = getNeuron(curr, index);
!
}
else if(name.equals("input")) {
!
String index = null;
double weight = -1.0;
!
for(int i=0; i<atts.getLength(); i++) {
Node att = atts.item(i);
!
String attName = att.getNodeName();
String attValue = att.getNodeValue();
!
if(attName.equals("ref")) {
index = new String(attValue);
}
else if(attName.equals("weight")) {
! try {
weight = Double.valueOf(attValue.trim()).doubleValue();
} catch (NumberFormatException nfe) {
--- 271,307 ----
}
else if(name.equals("neuron")) {
!
String index = new String();
!
for (int i = 0; i < atts.getLength(); i++) {
Node att = atts.item(i);
!
String attName = att.getNodeName();
String attValue = att.getNodeValue();
!
if(attName.equals("index")) {
index = new String(attValue);
}
}
!
currNeuron = getNeuron(curr, index);
!
}
else if(name.equals("input")) {
!
String index = null;
double weight = -1.0;
!
for(int i=0; i<atts.getLength(); i++) {
Node att = atts.item(i);
!
String attName = att.getNodeName();
String attValue = att.getNodeValue();
!
if(attName.equals("ref")) {
index = new String(attValue);
}
else if(attName.equals("weight")) {
! try {
weight = Double.valueOf(attValue.trim()).doubleValue();
} catch (NumberFormatException nfe) {
***************
*** 311,315 ****
}
}
! }
if(weight < 0.0) {
--- 309,313 ----
}
}
! }
if(weight < 0.0) {
***************
*** 320,326 ****
throw new ClassificationException("No \'ref\' value assigned for neuron " + currNeuron.name + " in layer " + currLayer);
}
!
System.out.println("Adding input " + index + " with weight " + weight);
!
if(currLayer > 0) {
Neuron toAdd = getNeuron((ArrayList)layers.get(currLayer-1), index);
--- 318,324 ----
throw new ClassificationException("No \'ref\' value assigned for neuron " + currNeuron.name + " in layer " + currLayer);
}
!
System.out.println("Adding input " + index + " with weight " + weight);
!
if(currLayer > 0) {
Neuron toAdd = getNeuron((ArrayList)layers.get(currLayer-1), index);
***************
*** 336,345 ****
}
else if(name.equals("output")) {
!
String index = null;
!
for(int i=0; i<atts.getLength(); i++) {
Node att = atts.item(i);
!
String attName = att.getNodeName();
String attValue = att.getNodeValue();
--- 334,343 ----
}
else if(name.equals("output")) {
!
String index = null;
!
for(int i=0; i<atts.getLength(); i++) {
Node att = atts.item(i);
!
String attName = att.getNodeName();
String attValue = att.getNodeValue();
***************
*** 348,359 ****
index = new String(attValue);
}
! }
!
if(index == null || index.equals("")) {
throw new ClassificationException("No \'ref\' value assigned for neuron " + currNeuron.name + " in layer " + currLayer);
}
!
System.out.println("Adding output " + index);
!
if(currLayer >= 0) {
Neuron toAdd = getNeuron((ArrayList)layers.get(currLayer+1), index);
--- 346,357 ----
index = new String(attValue);
}
! }
!
if(index == null || index.equals("")) {
throw new ClassificationException("No \'ref\' value assigned for neuron " + currNeuron.name + " in layer " + currLayer);
}
!
System.out.println("Adding output " + index);
!
if(currLayer >= 0) {
Neuron toAdd = getNeuron((ArrayList)layers.get(currLayer+1), index);
***************
*** 364,371 ****
currNeuron.addOutput(toAdd);
}
!
}
! }
!
// Recurse for children if any
for (Node child = n.getFirstChild(); child != null; child = child.getNextSibling()) {
--- 362,369 ----
currNeuron.addOutput(toAdd);
}
!
}
! }
!
// Recurse for children if any
for (Node child = n.getFirstChild(); child != null; child = child.getNextSibling()) {
***************
*** 383,387 ****
public void dumpXML() {
!
System.out.println("<?xml version=\"1.0\"?>");
System.out.println("<net>");
--- 381,385 ----
public void dumpXML() {
!
System.out.println("<?xml version=\"1.0\"?>");
System.out.println("<net>");
***************
*** 401,415 ****
System.out.println("\" index=\"" + i + "\"/>");
!
for(int j=0; j<tmp_layer.size(); j++) {
Neuron tmp_neuron = (Neuron)tmp_layer.get(j);
!
tmp_neuron.printXML(2);
}
!
indent(1);
System.out.println("</layer>");
}
!
System.out.println("</net>");
}
--- 399,413 ----
System.out.println("\" index=\"" + i + "\"/>");
!
for(int j=0; j<tmp_layer.size(); j++) {
Neuron tmp_neuron = (Neuron)tmp_layer.get(j);
!
tmp_neuron.printXML(2);
}
!
indent(1);
System.out.println("</layer>");
}
!
System.out.println("</net>");
}
***************
*** 433,437 ****
public void train(double [] in, double [] expected, double trainconst) throws ClassificationException
{
! if(trainconst <= 0.0)
throw new ClassificationException("Training constant must be >= 0.0");
--- 431,435 ----
public void train(double [] in, double [] expected, double trainconst) throws ClassificationException
{
! if(trainconst <= 0.0)
throw new ClassificationException("Training constant must be >= 0.0");
***************
*** 468,472 ****
ArrayList tmp = (ArrayList)layers.get(i);
! for(int j=0; j<tmp.size(); j++)
((Neuron)tmp.get(j)).commit();
}
--- 466,470 ----
ArrayList tmp = (ArrayList)layers.get(i);
! for(int j=0; j<tmp.size(); j++)
((Neuron)tmp.get(j)).commit();
}
***************
*** 475,490 ****
/* From Storage Manager */
!
! public void dump() throws IOException
! {
! throw new NotImplementedException("NeuralNetwork.dump()");
! }
!
public void restore() throws IOException
{
! throw new NotImplementedException("NeuralNetwork.restore()");
}
!
//iclement: This may need revision:
// Error handler to report errors and warnings
--- 473,490 ----
/* From Storage Manager */
!
! public void dump() throws IOException
! {
! //throw new NotImplementedException("NeuralNetwork.dump()");
! dumpXML();
! }
!
public void restore() throws IOException
{
! throw new NotImplementedException("NeuralNetwork.restore()");
}
!
//iclement: This may need revision:
+ //mokhov: i guess so
// Error handler to report errors and warnings
***************
*** 496,500 ****
this.out = out;
}
!
/**
* Returns a string describing parse exception details
--- 496,500 ----
this.out = out;
}
!
/**
* Returns a string describing parse exception details
***************
*** 517,526 ****
out.println("Warning: " + getParseExceptionInfo(spe));
}
!
public void error(SAXParseException spe) throws SAXException {
String message = "Error: " + getParseExceptionInfo(spe);
throw new SAXException(message);
}
!
public void fatalError(SAXParseException spe) throws SAXException {
String message = "Fatal Error: " + getParseExceptionInfo(spe);
--- 517,526 ----
out.println("Warning: " + getParseExceptionInfo(spe));
}
!
public void error(SAXParseException spe) throws SAXException {
String message = "Error: " + getParseExceptionInfo(spe);
throw new SAXException(message);
}
!
public void fatalError(SAXParseException spe) throws SAXException {
String message = "Fatal Error: " + getParseExceptionInfo(spe);
Index: Neuron.java
===================================================================
RCS file: /cvsroot/marf/marf/src/marf/Classification/NeuralNetwork/Neuron.java,v
retrieving revision 1.1
retrieving revision 1.2
diff -C2 -d -r1.1 -r1.2
*** Neuron.java 18 Nov 2002 01:27:36 -0000 1.1
--- Neuron.java 23 Nov 2002 16:46:56 -0000 1.2
***************
*** 1,2 ****
--- 1,4 ----
+ package marf.Classification.NeuralNetwork;
+
import java.util.ArrayList;
import java.lang.Double;
***************
*** 13,17 ****
private ArrayList weights = new ArrayList();
private ArrayList weightsBuf = new ArrayList();
!
private ArrayList outputs = new ArrayList();
--- 15,19 ----
private ArrayList weights = new ArrayList();
private ArrayList weightsBuf = new ArrayList();
!
private ArrayList outputs = new ArrayList();
***************
*** 28,33 ****
public boolean addInput(Neuron in, double weight) {
! return inputs.add(in) &&
! weights.add(new Double(weight)) &&
weightsBuf.add(new Double(weight));
}
--- 30,35 ----
public boolean addInput(Neuron in, double weight) {
! return inputs.add(in) &&
! weights.add(new Double(weight)) &&
weightsBuf.add(new Double(weight));
}
***************
*** 42,46 ****
if(inputs.isEmpty()) return;
!
double count = 0;
--- 44,48 ----
if(inputs.isEmpty()) return;
!
double count = 0;
***************
*** 54,66 ****
//System.out.println("Neuron: " + name + ", Sum: " + count + ", Result: " + result);
}
!
private double getWeight(Neuron n) {
int val = inputs.indexOf(n);
!
! if(val>=0)
return ((Double)weights.get(val)).doubleValue();
!
//System.out.println("There is no pointer n in neuron");
!
return -1.0;
}
--- 56,68 ----
//System.out.println("Neuron: " + name + ", Sum: " + count + ", Result: " + result);
}
!
private double getWeight(Neuron n) {
int val = inputs.indexOf(n);
!
! if(val>=0)
return ((Double)weights.get(val)).doubleValue();
!
//System.out.println("There is no pointer n in neuron");
!
return -1.0;
}
***************
*** 72,81 ****
}
else if(type == 1) {
!
double sum = 0.0;
!
for(int i=0; i<outputs.size(); i++)
sum += ((Neuron)outputs.get(i)).delta * ((Neuron)outputs.get(i)).getWeight(this);
!
delta = result * (1.0 - result) * sum;
}
--- 74,83 ----
}
else if(type == 1) {
!
double sum = 0.0;
!
for(int i=0; i<outputs.size(); i++)
sum += ((Neuron)outputs.get(i)).delta * ((Neuron)outputs.get(i)).getWeight(this);
!
delta = result * (1.0 - result) * sum;
}
***************
*** 87,91 ****
for(int i=0; i<inputs.size(); i++) {
! weightsBuf.set(i, new Double(beta * ((Double)weights.get(i)).doubleValue() +
alpha * delta * ((Neuron)inputs.get(i)).result));
--- 89,93 ----
for(int i=0; i<inputs.size(); i++) {
! weightsBuf.set(i, new Double(beta * ((Double)weights.get(i)).doubleValue() +
alpha * delta * ((Neuron)inputs.get(i)).result));
***************
*** 103,110 ****
public void printXML(int tab) {
!
indent(tab);
System.out.println("<neuron index=\"" + name + "\" thresh=\"" + threshold + "\">");
!
for(int i=0; i<inputs.size(); i++) {
indent(tab+1);
--- 105,112 ----
public void printXML(int tab) {
!
indent(tab);
System.out.println("<neuron index=\"" + name + "\" thresh=\"" + threshold + "\">");
!
for(int i=0; i<inputs.size(); i++) {
indent(tab+1);
|