org.lenskit.eval.traintest.predict

Class PredictEvalTask

java.lang.Object

org.lenskit.eval.traintest.predict.PredictEvalTask

All Implemented Interfaces:

EvalTask

public class PredictEvalTask
extends Object
implements EvalTask

An eval task that attempts to predict the user’s test ratings. It is generally a bad idea, as well as pointless, to have multiple predict tasks in a single experiment.

Constructor Summary

Constructors

Constructor and Description
PredictEvalTask() Create a new eval task.

Method Summary

Modifier and Type	Method and Description
void	addMetric(PredictMetric<?> metric) Add a prediction metric.
ConditionEvaluator	createConditionEvaluator(AlgorithmInstance algorithm, DataSet dataSet, Recommender rec) Set up a measurement of a single recommender.
void	finish() Finalize this eval task.
static PredictEvalTask	fromJSON(com.fasterxml.jackson.databind.JsonNode json, URI base) Create a predict eval task from a JSON/YAML file.
List<Metric<?>>	getAllMetrics() Get the list of all metrics.
List<String>	getGlobalColumns() Get columns that will go in the aggregate output file.
Path	getOutputFile() Get the output file for writing predictions.
List<PredictMetric<?>>	getPredictMetrics() Get the list of prediction metrics.
Set<Class<?>>	getRequiredRoots() Get the root types required by this evaluation.
List<String>	getUserColumns() Get columns that will go in the per-user output file.
void	setOutputFile(Path file) Set the output file for predictions.
void	start(ExperimentOutputLayout outputLayout) Do initial setup for this eval task.

Methods inherited from class java.lang.Object

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

Constructor Detail

PredictEvalTask

public PredictEvalTask()

Create a new eval task.

Method Detail

fromJSON

public static PredictEvalTask fromJSON(com.fasterxml.jackson.databind.JsonNode json,
                                       URI base)
                                throws IOException

Create a predict eval task from a JSON/YAML file.

Parameters:

json - The task specification.

base - The base URI from which json came, used for resolving relative paths.

Returns:

The task.

Throws:

IOException

getOutputFile

public Path getOutputFile()

Get the output file for writing predictions.

Returns:

The output file, or null if no file is configured.

setOutputFile

public void setOutputFile(Path file)

Set the output file for predictions.

Parameters:

file - The output file for writing predictions. Will get a CSV file.

getPredictMetrics

public List<PredictMetric<?>> getPredictMetrics()

Get the list of prediction metrics.

Returns:

The list of prediction metrics. This list is live, not copied, so it can be modified or cleared.

getAllMetrics

public List<Metric<?>> getAllMetrics()

Get the list of all metrics.

Returns:

A list containing all metrics used by this task.

addMetric

public void addMetric(PredictMetric<?> metric)

Add a prediction metric.

Parameters:

metric - The metric to add.

getRequiredRoots

public Set<Class<?>> getRequiredRoots()

Description copied from interface: EvalTask

Get the root types required by this evaluation.

Specified by:

getRequiredRoots in interface EvalTask

Returns:

The root types required by this evaluation.

getGlobalColumns

public List<String> getGlobalColumns()

Description copied from interface: EvalTask

Get columns that will go in the aggregate output file.

Specified by:

getGlobalColumns in interface EvalTask

Returns:

The list of column names that this task will contribute to the aggregate output file.

getUserColumns

public List<String> getUserColumns()

Description copied from interface: EvalTask

Get columns that will go in the per-user output file.

Specified by:

getUserColumns in interface EvalTask

Returns:

The list of column names that this task will contribute to the per-user output file.

start

public void start(ExperimentOutputLayout outputLayout)

Description copied from interface: EvalTask

Do initial setup for this eval task. This should create any per-task output files, etc.

Specified by:

start in interface EvalTask

Parameters:

outputLayout - The output layout for experiment results.

finish

public void finish()

Description copied from interface: EvalTask

Finalize this eval task. This should finish writing and close any per-task output files, etc.

Specified by:

finish in interface EvalTask

createConditionEvaluator

public ConditionEvaluator createConditionEvaluator(AlgorithmInstance algorithm,
                                                   DataSet dataSet,
                                                   Recommender rec)

Description copied from interface: EvalTask

Set up a measurement of a single recommender.

Specified by:

createConditionEvaluator in interface EvalTask

Parameters:

algorithm - The algorithm being evaluated.

dataSet - The data set being evaluated.

rec - The recommender to measure.

Returns:

A condition evaluator that will measure the recommender’s performance on the algorithm and data set.