/*
* Encog(tm) Core v3.3 - Java Version
* http://www.heatonresearch.com/encog/
* https://github.com/encog/encog-java-core
* Copyright 2008-2014 Heaton Research, Inc.
*
* 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.
*
* For more information on Heaton Research copyrights, licenses
* and trademarks visit:
* http://www.heatonresearch.com/copyright
*/
package org.encog.ml.data.versatile.normalizers.strategies;
import java.util.HashMap;
import java.util.Map;
import org.encog.EncogError;
import org.encog.ml.data.MLData;
import org.encog.ml.data.versatile.columns.ColumnDefinition;
import org.encog.ml.data.versatile.columns.ColumnType;
import org.encog.ml.data.versatile.normalizers.Normalizer;
import org.encog.ml.data.versatile.normalizers.OneOfNNormalizer;
import org.encog.ml.data.versatile.normalizers.RangeNormalizer;
import org.encog.ml.data.versatile.normalizers.RangeOrdinal;
/**
* Provides a basic normalization strategy that will work with most models built into Encog.
* This is often used as a starting point for building more customized models, as this
* normalizer works mainly by using maps to define which normalizer to use for what
* data type.
*/
public class BasicNormalizationStrategy implements NormalizationStrategy {
/**
* Mapping to all of the input normalizers.
*/
private final Map<ColumnType,Normalizer> inputNormalizers = new HashMap<ColumnType,Normalizer>();
/**
* Mapping to all of the output normalizers.
*/
private final Map<ColumnType,Normalizer> outputNormalizers = new HashMap<ColumnType,Normalizer>();
/**
* Construct the basic normalization strategy.
* @param inputLow The desired low to normalize input into.
* @param inputHigh The desired high to normalize input into.
* @param outputLow The desired low to normalize output into.
* @param outputHigh The desired high to normalize output into.
*/
public BasicNormalizationStrategy(double inputLow, double inputHigh, double outputLow, double outputHigh) {
assignInputNormalizer(ColumnType.continuous,new RangeNormalizer(inputLow,inputHigh));
assignInputNormalizer(ColumnType.nominal,new OneOfNNormalizer(inputLow,inputHigh));
assignInputNormalizer(ColumnType.ordinal,new RangeOrdinal(inputLow,inputHigh));
assignOutputNormalizer(ColumnType.continuous,new RangeNormalizer(outputLow,outputHigh));
assignOutputNormalizer(ColumnType.nominal,new OneOfNNormalizer(outputLow,outputHigh));
assignOutputNormalizer(ColumnType.ordinal,new RangeOrdinal(outputLow,outputHigh));
}
/**
* Default constructor.
*/
public BasicNormalizationStrategy() {
}
/**
* Assign a normalizer to the specified column type for output.
* @param colType The column type.
* @param norm The normalizer.
*/
public void assignInputNormalizer(ColumnType colType, Normalizer norm) {
this.inputNormalizers.put(colType, norm);
}
/**
* Assign a normalizer to the specified column type for output.
* @param colType The column type.
* @param norm The normalizer.
*/
public void assignOutputNormalizer(ColumnType colType, Normalizer norm) {
this.outputNormalizers.put(colType, norm);
}
/**
* Find a normalizer for the specified column definition, and if it is input or output.
* @param colDef The column definition.
* @param isInput True if the column is input.
* @return The normalizer to use.
*/
private Normalizer findNormalizer(ColumnDefinition colDef, boolean isInput) {
Normalizer norm = null;
if(isInput) {
if( this.inputNormalizers.containsKey(colDef.getDataType())) {
norm = this.inputNormalizers.get(colDef.getDataType());
}
} else {
if( this.outputNormalizers.containsKey(colDef.getDataType())) {
norm = this.outputNormalizers.get(colDef.getDataType());
}
}
if( norm==null ) {
throw new EncogError("No normalizer defined for input="+isInput+", type=" + colDef.getDataType());
}
return norm;
}
/**
* {@inheritDoc}
*/
@Override
public int normalizedSize(ColumnDefinition colDef, boolean isInput) {
Normalizer norm = findNormalizer(colDef,isInput);
return norm.outputSize(colDef);
}
/**
* {@inheritDoc}
*/
@Override
public int normalizeColumn(ColumnDefinition colDef, boolean isInput,
String value, double[] outputData, int outputColumn) {
Normalizer norm = findNormalizer(colDef,isInput);
return norm.normalizeColumn(colDef,value,outputData,outputColumn);
}
/**
* {@inheritDoc}
*/
@Override
public int normalizeColumn(ColumnDefinition colDef, boolean isInput,
double value, double[] outputData, int outputColumn) {
Normalizer norm = findNormalizer(colDef,isInput);
return norm.normalizeColumn(colDef,value,outputData,outputColumn);
}
/**
* @return the inputNormalizers
*/
public Map<ColumnType, Normalizer> getInputNormalizers() {
return inputNormalizers;
}
/**
* @return the outputNormalizers
*/
public Map<ColumnType, Normalizer> getOutputNormalizers() {
return outputNormalizers;
}
/**
* {@inheritDoc}
*/
@Override
public String denormalizeColumn(ColumnDefinition colDef, boolean isInput,
MLData data, int dataColumn) {
Normalizer norm = findNormalizer(colDef,isInput);
return norm.denormalizeColumn(colDef,data,dataColumn);
}
}