MATLAB File Help: prtOutlierRemovalNStd |
prtOutlierRemovalNStd Removes outliers from a prtDataSet NSTDOUT = prtOutlierRemovalNStd creates a pre-processing object that flags as outliers data where any of the feature values is more then nStd standard deviations from the mean of that feature. prtOutlierRemovalNStd has the following properties: nStd - The number of standard deviations at which to flag an observation as an outlier an observation (default = 3) A prtOutlierRemovalNStd object also inherits all properties and functions from the prtOutlierRemoval class. For more information on how to control the behaviour of outlier removal objects, see the help for prtOutlierRemoval. Example: dataSet = prtDataGenUnimodal; % Load a data Set outlier = prtDataSetClass([-10 -10],1); % Create and insert dataSet = catObservations(dataSet,outlier); % an outlier % Create the prtOutlierRemoval object nStdRemove = prtOutlierRemovalNStd('runMode','removeObservation'); nStdRemove = nStdRemove.train(dataSet); % Train and run dataSetNew = nStdRemove.run(dataSet); % Plot the results subplot(2,1,1); plot(dataSet); title('Original Data'); subplot(2,1,2); plot(dataSetNew); title('NstdOutlierRemove Data');
Superclasses | prtOutlierRemoval |
Sealed | false |
Construct on load | false |
prtOutlierRemovalNStd | Removes outliers from a prtDataSet |
dataSet | The training prtDataSet, only stored if verboseStorage is true. |
dataSetSummary | Structure that summarizes prtDataSet. |
isCrossValidateValid | False |
isSupervised | False |
isTrained | Indicates if prtAction object has been trained. |
meanVector | The mean vector |
nStd | The number of standard deviations beyond which to remove data |
name | Standard Deviation Based Outlier Removal |
nameAbbreviation | nStd |
replaceValue | |
runMode | Operation taken during RUN |
runOnTrainingMode | Operation taken during TRAIN |
showProgressBar | |
stdVector | The standard deviation vector |
userData | User specified data |
validModes | |
verboseStorage | Specifies whether or not to store the training prtDataset. |
crossValidate | Cross validate prtAction using prtDataSet and cross validation keys. | |
get | get the object properties | |
kfolds | Perform K-folds cross-validation of prtAction | |
optimize | Optimize action parameter by exhaustive function maximization. | |
run | Run a prtAction object on a prtDataSet object. | |
set | set the object properties | |
train | Train a prtAction object using training a prtDataSet object. |