| AbstractRowFinderApplier |
Ancestor for filters that apply RowFinder schemes to the data. |
| AccumulatedLWLWeights |
Generates an LWL-like dataset for each instance of the data from the first batch and accumulate these weights.
|
| AlignDataset |
Aligns the dataset(s) passing through to the reference dataset.
Makes use of the following other filters internally:
- weka.filters.unsupervised.attribute.AnyToString
- weka.filters.unsupervised.instance.RemoveWithLabels
Valid options are:
|
| DatasetCleaner |
Removes all rows from the data data that have been indentified.
|
| DatasetLabeler |
Adds an additional attribute to the dataset containing a label whether it was a match or not, i.e., whether the row finder selected a particular row or not.
|
| KeepRange |
Keeps only the range of rows, in the order specified.
|
| KennardStone |
Applies the Kennard-Stone algorithm to the dataset.
Each row has the pre-filter (eg PLS) applied before performing the search.
|
| LatestRecords |
Retains the latest database records.
|
| MultiRowProcessor |
Uses the specified row selection scheme to identify groups of rows in the data coming through and then applies the selected row processor to these subsets.
|
| RemoveDuplicateIDs |
Removes rows with IDs that occur multiple times.
Also skips rows with missing ID.
|
| RemoveDuplicates |
Removes all duplicate instances.
|
| RemoveInstancesWithMissingValue |
Removes all instances that contain missing values.
|
| RemoveMisclassifiedAbs |
A filter that removes instances which are incorrectly classified.
|
| RemoveMisclassifiedRel |
A filter that removes instances which are incorrectly classified.
|
| RemoveTestInstances |
Removes all instances of the provided test set from the data passing through.
Requires an attribute in the data that uniquely identifies instances across datasets.
|
| RemoveWithLabels |
Allows the user to remove nominal labels via a regular expression.
|
| RemoveWithWeights |
Removes instances with weights outside the defined limits.
|
| RemoveWithZeroes |
Removes all instances that contain at least the specified number (or percentage) of zeroes in numeric attributes.
|
| RowNorm |
Row wise normalization.
|
| SafeRemoveRange |
A filter that removes a given range of instances of a dataset.
Works just like weka.filters.unsupervised.instance.RemoveRange, but has a more robust handling of instance ranges.
|
| Scale |
Scales all numeric attributes between the specified min/max.
|
| Sort |
Sorts the instances.
|
| SortOnAttribute |
Sorts the instances on a particular attribute.
|
| WeightsBasedResample |
Normalizes all instance weights and drops the ones that fall below the specified threshold, but at most the specified percentage.
Of the left over instances, the smallest weight, e.g., 0.2, represents one instance, which translates a weight of 1.0 to five instances.
|