Package ubic.gemma.core.analysis.preprocess

Interface OutlierDetectionService

All Known Implementing Classes:

OutlierDetectionServiceImpl

public interface OutlierDetectionService

Author:

paul

Method Summary

Modifier and Type	Method	Description
Optional<Collection<OutlierDetails>>	getOutlierDetails(ExpressionExperiment ee)	Obtain existing outlier details from a previous identifyOutliersByMedianCorrelation(ExpressionExperiment) invocation.
Collection<OutlierDetails>	identifyOutliersByMedianCorrelation(DoubleMatrix<BioAssay,BioAssay> cormat)	Identify outliers by sorting by median, then looking for non-overlap of first quartile-second quartile range This is exposed for efficiency in geeq score calculation, use this#identifyOutliers(ExpressionExperiment, boolean, boolean) to have the correlation matrix computed correctly for you.
Collection<OutlierDetails>	identifyOutliersByMedianCorrelation(ExpressionExperiment ee)	Identify outliers by median correlation for the given dataset.

Method Details

getOutlierDetails

Optional<Collection<OutlierDetails>> getOutlierDetails(ExpressionExperiment ee)

Obtain existing outlier details from a previous identifyOutliersByMedianCorrelation(ExpressionExperiment) invocation.

Returns:

outlier details or Optional#empty() if no sample coexpression analysis exists

identifyOutliersByMedianCorrelation

Collection<OutlierDetails> identifyOutliersByMedianCorrelation(ExpressionExperiment ee)
                                                        throws FilteringException

Identify outliers by median correlation for the given dataset.

Parameters:

ee - The experiment to identify sample outliers in.

Returns:

the information about the identified outliers.

Throws:

FilteringException

See Also:

• identifyOutliersByMedianCorrelation(DoubleMatrix)

identifyOutliersByMedianCorrelation

Collection<OutlierDetails> identifyOutliersByMedianCorrelation(DoubleMatrix<BioAssay,BioAssay> cormat)

Identify outliers by sorting by median, then looking for non-overlap of first quartile-second quartile range This is exposed for efficiency in geeq score calculation, use this#identifyOutliers(ExpressionExperiment, boolean, boolean) to have the correlation matrix computed correctly for you.

Parameters:

cormat - the correlation matrix to identify outliers in.

Returns:

the information about the identified outliers.