Package de.fu_berlin.ties.eval

This packages provides functionality for evaluating results of classifiers and extractors.

See:
          Description

Interface Summary
EvalInput Classes implementing this interface provide input for calculating evaluation metrics: true positives, false negatives and false positives.
FeatureCountView Provides a read-only view on the statistics calculated by the FeatureCount class and the underlying raw counts.
FMetricsSummary Implementations of this interface can show statistical summaries of precision, recall, and F1 metrics updated in several operations.
FMetricsView A read-only view of the evaluation results calculated by the FMetrics class and the underlying raw counts.
MultiFMetricsView A read-only view of multiple FMetrics and the sums and averages calculated over them.
 

Class Summary
Accuracy Counts true and false items and measures the accuracy: A = true / (true + false).
EvalStatus Type-safe enumeration of possible evaluation states for predictions (unknown, correct, spurious etc.) and answer keys (missing etc.).
FeatureCount Keeps track of the average number of features and of unique features in context representations and of the average number of contexts in documents.
FMetrics This class manages and updates evaluation results, calculating precision (P), recall (R) and F-measure (F).
MultiFMetrics Instances of this class manage multiple FMetrics for different types.
ReEvaluator A processor that can be used to re-evaluate the contents of an EvaluatedExtractionContainer.
SummaryFMetrics FMetrics extension that additionally calculates a StatisticalSummary of the intermediate precision, recall, and F1 metrics resulting from different update operations.
 

Package de.fu_berlin.ties.eval Description

This packages provides functionality for evaluating results of classifiers and extractors.



Copyright © 2003-2004 Christian Siefkes. All Rights Reserved.