meka.core

Class Result

java.lang.Object

meka.core.Result

All Implemented Interfaces:

java.io.Serializable

public class Result
extends java.lang.Object
implements java.io.Serializable

Result - Stores predictions alongside true labels, for evaluation. For more on the evaluation and threshold selection implemented here; see:

Jesse Read, Bernhard Pfahringer, Geoff Holmes, Eibe Frank. Classifier Chains for Multi-label Classification. Machine Learning Journal. Springer (2011).
Jesse Read, Scalable Multi-label Classification. PhD Thesis, University of Waikato, Hamilton, New Zealand (2010).

Version:

March 2012 - Multi-target Compatible

Author:

Jesse Read

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field and Description
java.util.ArrayList<int[]>	actuals
java.util.HashMap<java.lang.String,java.lang.String>	info
int	L
java.util.HashMap<java.lang.String,java.lang.Double>	output
java.util.ArrayList<double[]>	predictions
java.util.HashMap<java.lang.String,java.lang.Double>	vals

Constructor Summary

Constructors

Constructor and Description
Result()
Result(int L)
Result(int N, int L)

Method Summary

Methods

Modifier and Type	Method and Description
void	addResult(double[] pred, weka.core.Instance real) AddResult - Add an entry.
void	addValue(java.lang.String metric, double v) AddValue.
int[][]	allActuals() AllActuals - Retrive all true predictions in an L x N matrix.
double[][]	allPredictions() AllPredictions - Retrive all prediction confidences in an L * N matrix.
int[][]	allPredictions(double t) AllPredictions - Retrive all predictions (according to threshold t) in an L * N matrix.
java.lang.String	getInfo(java.lang.String cat) GetInfo.
static java.lang.String	getResultAsString(Result s) GetResultAsString - print out each prediction in a Result along with its true labelset.
static java.lang.String	getResultAsString(Result s, int adp) GetResultAsString - print out each prediction in a Result (to a certain number of decimal points) along with its true labelset.
static java.util.HashMap<java.lang.String,java.lang.Double>	getStats(Result r, java.lang.String vop) GetStats.
double	getValue(java.lang.String metric) AddValue.
int[]	rowActual(int i) RowActual - Retrive the true values for the i-th instance.
int[]	rowPrediction(int i) RowPrediction - Retrive the predicted values for the i-th instance according to pre-calibrated/chosen threshold.
int[]	rowPrediction(int i, double t) RowPrediction - Retrive the predicted values for the i-th instance according to threshold t.
double[]	rowRanking(int i) RowRanking - Retrive the prediction confidences for the i-th instance.
void	setInfo(java.lang.String cat, java.lang.String val) SetInfo.
void	setValue(java.lang.String metric, double v) SetValue.
int	size()
java.lang.String	toString()
static void	writeResultToFile(Result result, java.lang.String fname) WriteResultToFile -- write a Result 'result' out in plain text format to file 'fname'.

Methods inherited from class java.lang.Object

equals, getClass, hashCode, notify, notifyAll, wait, wait, wait

Field Detail

L

public int L

predictions

public java.util.ArrayList<double[]> predictions

actuals

public java.util.ArrayList<int[]> actuals

output

public java.util.HashMap<java.lang.String,java.lang.Double> output

info

public java.util.HashMap<java.lang.String,java.lang.String> info

vals

public java.util.HashMap<java.lang.String,java.lang.Double> vals

Constructor Detail

Result

public Result()

Result

public Result(int L)

Result

public Result(int N,
      int L)

Method Detail

size

public int size()

toString

public java.lang.String toString()

Overrides:

toString in class java.lang.Object

addResult

public void addResult(double[] pred,
             weka.core.Instance real)

AddResult - Add an entry.

rowActual

public int[] rowActual(int i)

RowActual - Retrive the true values for the i-th instance.

rowRanking

public double[] rowRanking(int i)

RowRanking - Retrive the prediction confidences for the i-th instance.
TODO rename to rowConfidence

rowPrediction

public int[] rowPrediction(int i,
                  double t)

RowPrediction - Retrive the predicted values for the i-th instance according to threshold t.

rowPrediction

public int[] rowPrediction(int i)

RowPrediction - Retrive the predicted values for the i-th instance according to pre-calibrated/chosen threshold.

allPredictions

public double[][] allPredictions()

AllPredictions - Retrive all prediction confidences in an L * N matrix.

allPredictions

public int[][] allPredictions(double t)

AllPredictions - Retrive all predictions (according to threshold t) in an L * N matrix.

allActuals

public int[][] allActuals()

AllActuals - Retrive all true predictions in an L x N matrix.

addValue

public void addValue(java.lang.String metric,
            double v)

AddValue. Add v to an existing metric value.

setValue

public void setValue(java.lang.String metric,
            double v)

SetValue. Add an evaluation metric and a value for it.

getValue

public double getValue(java.lang.String metric)

AddValue. Retrieve the value for metric 'metric'

setInfo

public void setInfo(java.lang.String cat,
           java.lang.String val)

SetInfo. Set a String value to an information category.

getInfo

public java.lang.String getInfo(java.lang.String cat)

GetInfo. Get the String value of category 'cat'.

getStats

public static java.util.HashMap<java.lang.String,java.lang.Double> getStats(Result r,
                                                            java.lang.String vop)

GetStats. Return the evaluation statistics given predictions and real values stored in r. In the multi-label case, a Threshold category must exist, containing a string defining the type of threshold we want to use/calibrate.

getResultAsString

public static java.lang.String getResultAsString(Result s)

GetResultAsString - print out each prediction in a Result along with its true labelset.

writeResultToFile

public static void writeResultToFile(Result result,
                     java.lang.String fname)
                              throws java.lang.Exception

WriteResultToFile -- write a Result 'result' out in plain text format to file 'fname'.

Parameters:

result - Result

fname - file name

Throws:

java.lang.Exception

getResultAsString

public static java.lang.String getResultAsString(Result s,
                                 int adp)

GetResultAsString - print out each prediction in a Result (to a certain number of decimal points) along with its true labelset.