org.apache.ctakes.ytex.kernel

Class BaseClassifierEvaluationParser

java.lang.Object

org.apache.ctakes.ytex.kernel.BaseClassifierEvaluationParser

All Implemented Interfaces:

ClassifierEvaluationParser

Direct Known Subclasses:

LibSVMParser, SemiLEvaluationParser, SvmlinEvaluationParser

public abstract class BaseClassifierEvaluationParser
extends Object
implements ClassifierEvaluationParser

miscellaneous methods used for parsing various output types

Author:

vhacongarlav

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	BaseClassifierEvaluationParser.InstanceClassInfo

Nested classes/interfaces inherited from interface org.apache.ctakes.ytex.kernel.ClassifierEvaluationParser

ClassifierEvaluationParser.ParseOption

Field Summary

Fields

Modifier and Type	Field and Description
static Pattern	wsDotPattern
static Pattern	wsPattern

Fields inherited from interface org.apache.ctakes.ytex.kernel.ClassifierEvaluationParser

NO, YES

Constructor Summary

Constructors

Constructor and Description
BaseClassifierEvaluationParser()

Method Summary

Modifier and Type	Method and Description
protected boolean	checkFileRead(String file)
static String	extractFirstToken(String line, Pattern tokDelimPattern)
ClassifierEvaluationDao	getClassifierEvaluationDao()
protected String	getFileBaseName(Properties kernelProps)
protected void	initClassifierEvaluation(String instanceIdFile, ClassifierEvaluation eval)
protected void	initClassifierEvaluationFromProperties(Properties props, ClassifierEvaluation eval)
protected com.google.common.collect.BiMap<Integer,String>	loadClassIdMap(File dataDir, String label)
protected List<List<Long>>	loadClassInfo(File dataDir, String classFileName)
protected List<BaseClassifierEvaluationParser.InstanceClassInfo>	loadInstanceClassInfo(File dataDir, String classFileName)
Properties	loadProps(File outputDir) load properties from outputDir/options.properties.
protected Double	parseDoubleOption(Pattern pCost, String options) parse a number out of the libsvm command line that matches the specified pattern.
List<Long>	parseInstanceIds(String instanceIdFile)
protected Integer	parseIntOption(Pattern pKernel, String options) parse a number out of the libsvm command line that matches the specified pattern.
void	setClassifierEvaluationDao(ClassifierEvaluationDao classifierEvaluationDao)
protected void	storeSemiSupervised(Properties kernelProps, ClassifierEvaluation ce, com.google.common.collect.BiMap<Integer,String> classIdToNameMap)
protected void	updateSemiSupervisedPredictions(ClassifierEvaluation ce, List<BaseClassifierEvaluationParser.InstanceClassInfo> listClassInfo, boolean storeUnlabeled, String[] predictedClassNames, Map<String,Integer> classNameToIdMap)
protected void	updateSemiSupervisedPredictions(ClassifierEvaluation ce, List<List<Long>> listClassInfo, boolean storeUnlabeled, int[] classIds) used by semil & svmlin to store semisupervised predictions.

Methods inherited from class java.lang.Object

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

Methods inherited from interface org.apache.ctakes.ytex.kernel.ClassifierEvaluationParser

parseDirectory

Field Detail

wsPattern

public static Pattern wsPattern

wsDotPattern

public static Pattern wsDotPattern

Constructor Detail

BaseClassifierEvaluationParser

public BaseClassifierEvaluationParser()

Method Detail

getClassifierEvaluationDao

public ClassifierEvaluationDao getClassifierEvaluationDao()

setClassifierEvaluationDao

public void setClassifierEvaluationDao(ClassifierEvaluationDao classifierEvaluationDao)

extractFirstToken

public static String extractFirstToken(String line,
                                       Pattern tokDelimPattern)

parseInstanceIds

public List<Long> parseInstanceIds(String instanceIdFile)
                            throws IOException

Throws:

IOException

parseDoubleOption

protected Double parseDoubleOption(Pattern pCost,
                                   String options)

parse a number out of the libsvm command line that matches the specified pattern.

Parameters:

pCost -

options -

Returns:

null if option not present

parseIntOption

protected Integer parseIntOption(Pattern pKernel,
                                 String options)

parse a number out of the libsvm command line that matches the specified pattern.

Parameters:

pKernel -

options -

Returns:

null if option not present

initClassifierEvaluation

protected void initClassifierEvaluation(String instanceIdFile,
                                        ClassifierEvaluation eval)

initClassifierEvaluationFromProperties

protected void initClassifierEvaluationFromProperties(Properties props,
                                                      ClassifierEvaluation eval)

loadProps

public Properties loadProps(File outputDir)
                     throws FileNotFoundException,
                            IOException

load properties from outputDir/options.properties. returns empty properties if the file does not exist

Parameters:

outputDir -

Returns:

Throws:

FileNotFoundException

IOException

checkFileRead

protected boolean checkFileRead(String file)

getFileBaseName

protected String getFileBaseName(Properties kernelProps)

storeSemiSupervised

protected void storeSemiSupervised(Properties kernelProps,
                                   ClassifierEvaluation ce,
                                   com.google.common.collect.BiMap<Integer,String> classIdToNameMap)

updateSemiSupervisedPredictions

protected void updateSemiSupervisedPredictions(ClassifierEvaluation ce,
                                               List<List<Long>> listClassInfo,
                                               boolean storeUnlabeled,
                                               int[] classIds)

used by semil & svmlin to store semisupervised predictions. these train ml and make test predictions in a single step.

Parameters:

ce - updated

listClassInfo - the class info 0 - instance id, 1 - train/test, 2 - target class id

storeUnlabeled - should the unlabeled predictions be stored?

classIds - predicted class ids

updateSemiSupervisedPredictions

protected void updateSemiSupervisedPredictions(ClassifierEvaluation ce,
                                               List<BaseClassifierEvaluationParser.InstanceClassInfo> listClassInfo,
                                               boolean storeUnlabeled,
                                               String[] predictedClassNames,
                                               Map<String,Integer> classNameToIdMap)

loadClassIdMap

protected com.google.common.collect.BiMap<Integer,String> loadClassIdMap(File dataDir,
                                                                         String label)
                                                                  throws IOException

Throws:

IOException

loadInstanceClassInfo

protected List<BaseClassifierEvaluationParser.InstanceClassInfo> loadInstanceClassInfo(File dataDir,
                                                                                       String classFileName)
                                                                                throws IOException

Throws:

IOException

loadClassInfo

protected List<List<Long>> loadClassInfo(File dataDir,
                                         String classFileName)
                                  throws IOException

Throws:

IOException