Package | Description |
---|---|
de.lmu.ifi.dbs.elki.datasource |
Data normalization (and reconstitution) of data sets
|
de.lmu.ifi.dbs.elki.datasource.bundle |
Object bundles - exchange container for multi-represented objects
|
de.lmu.ifi.dbs.elki.datasource.filter |
Data filtering, in particular for normalization and projection
|
de.lmu.ifi.dbs.elki.datasource.filter.cleaning |
Filters for data cleaning.
|
de.lmu.ifi.dbs.elki.datasource.filter.normalization.instancewise |
Instancewise normalization, where each instance is normalized independently.
|
de.lmu.ifi.dbs.elki.datasource.filter.selection |
Filters for selecting and sorting data to process.
|
de.lmu.ifi.dbs.elki.datasource.filter.transform |
Data space transformations
|
de.lmu.ifi.dbs.elki.datasource.filter.typeconversions |
Filters to perform data type conversions.
|
de.lmu.ifi.dbs.elki.datasource.parser |
Parsers for different file formats and data types
The general use-case for any parser is to create objects out of an
InputStream (e.g. by reading a data file). |
Modifier and Type | Method and Description |
---|---|
protected BundleStreamSource |
AbstractDatabaseConnection.invokeStreamFilters(BundleStreamSource stream)
Transforms the specified list of objects and their labels into a list of
objects and their associations.
|
Modifier and Type | Method and Description |
---|---|
protected BundleStreamSource |
AbstractDatabaseConnection.invokeStreamFilters(BundleStreamSource stream)
Transforms the specified list of objects and their labels into a list of
objects and their associations.
|
Modifier and Type | Class and Description |
---|---|
class |
BundleReader
Read an ELKI bundle file into a data stream.
|
class |
StreamFromBundle
Convert a MultipleObjectsBundle to a stream.
|
Modifier and Type | Method and Description |
---|---|
BundleStreamSource |
MultipleObjectsBundle.asStream()
Process this bundle as stream.
|
Modifier and Type | Method and Description |
---|---|
static MultipleObjectsBundle |
MultipleObjectsBundle.fromStream(BundleStreamSource source)
Convert an object stream to a bundle
|
void |
BundleWriter.writeBundleStream(BundleStreamSource source,
java.nio.channels.WritableByteChannel output)
Write a bundle stream to a file output channel.
|
private ByteBufferSerializer<?>[] |
BundleWriter.writeHeader(BundleStreamSource source,
java.nio.ByteBuffer buffer,
java.nio.channels.WritableByteChannel output)
Write the header for the given stream to the stream.
|
Modifier and Type | Interface and Description |
---|---|
interface |
StreamFilter
Streaming filters are often more efficient (less memory use) as they do not
keep a reference to earlier data.
|
Modifier and Type | Class and Description |
---|---|
class |
AbstractStreamConversionFilter<I,O>
Abstract base class for simple conversion filters such as normalizations and
projections.
|
class |
AbstractStreamFilter
Abstract base class for streaming filters.
|
class |
AbstractVectorStreamConversionFilter<I,O extends NumberVector>
Abstract base class for streaming filters that produce vectors.
|
class |
NoOpFilter
Dummy filter that doesn't do any filtering.
|
Modifier and Type | Field and Description |
---|---|
protected BundleStreamSource |
AbstractStreamFilter.source
Data source
|
Modifier and Type | Method and Description |
---|---|
BundleStreamSource |
StreamFilter.init(BundleStreamSource source)
Connect to the previous stream.
|
BundleStreamSource |
AbstractStreamFilter.init(BundleStreamSource source) |
Modifier and Type | Method and Description |
---|---|
BundleStreamSource |
StreamFilter.init(BundleStreamSource source)
Connect to the previous stream.
|
BundleStreamSource |
AbstractStreamFilter.init(BundleStreamSource source) |
Modifier and Type | Class and Description |
---|---|
class |
DropNaNFilter
A filter to drop all records that contain NaN values.
|
class |
NoMissingValuesFilter
A filter to remove entries that have missing values.
|
class |
ReplaceNaNWithRandomFilter
A filter to replace all NaN values with random values.
|
class |
VectorDimensionalityFilter<V extends NumberVector>
Filter to remove all vectors that do not have the desired dimensionality.
|
Modifier and Type | Class and Description |
---|---|
class |
HellingerHistogramNormalization<V extends NumberVector>
Normalize histograms by scaling them to unit absolute sum, then taking the
square root of the absolute value in each attribute, times the normalization
constant \(1/\sqrt{2}\).
|
class |
InstanceLogRankNormalization<V extends NumberVector>
Normalize vectors such that the smallest value of each instance is 0, the
largest is 1, but using \( \log_2(1+x) \).
|
class |
InstanceMeanVarianceNormalization<V extends NumberVector>
Normalize vectors such that they have zero mean and unit variance.
|
class |
InstanceMinMaxNormalization<V extends NumberVector>
Normalize vectors with respect to a given minimum and maximum in each
dimension.
|
class |
InstanceRankNormalization<V extends NumberVector>
Normalize vectors such that the smallest value of each instance is 0, the
largest is 1.
|
class |
LengthNormalization<V extends NumberVector>
Class to perform a normalization on vectors to norm 1.
|
class |
Log1PlusNormalization<V extends NumberVector>
Normalize the data set by applying \( \frac{\log(1+|x|b)}{\log 1+b} \) to any
value.
|
Modifier and Type | Class and Description |
---|---|
class |
ByLabelFilter
A filter to select data set by their label.
|
class |
FirstNStreamFilter
Keep only the first N elements of the data source.
|
class |
RandomSamplingStreamFilter
Subsampling stream filter.
|
Modifier and Type | Class and Description |
---|---|
class |
HistogramJitterFilter<V extends NumberVector>
Add Jitter, preserving the histogram properties (same sum, nonnegative).
|
class |
LatLngToECEFFilter<V extends NumberVector>
Project a 2D data set (latitude, longitude) to a 3D coordinate system (X, Y,
Z), such that Euclidean distance is line-of-sight.
|
class |
LngLatToECEFFilter<V extends NumberVector>
Project a 2D data set (longitude, latitude) to a 3D coordinate system (X, Y,
Z), such that Euclidean distance is line-of-sight.
|
class |
NumberVectorFeatureSelectionFilter<V extends NumberVector>
Parser to project the ParsingResult obtained by a suitable base parser onto a
selected subset of attributes.
|
class |
NumberVectorRandomFeatureSelectionFilter<V extends NumberVector>
Parser to project the ParsingResult obtained by a suitable base parser onto a
randomly selected subset of attributes.
|
class |
ProjectionFilter<I,O>
Apply a projection to the data.
|
Modifier and Type | Class and Description |
---|---|
class |
ClassLabelFromPatternFilter
Streaming filter to derive an outlier class label.
|
class |
MultivariateTimeSeriesFilter<V extends FeatureVector<?>>
Class to "fold" a flat number vector into a multivariate time series.
|
Modifier and Type | Interface and Description |
---|---|
interface |
StreamingParser
Interface for streaming parsers, that may be much more efficient in
combination with filters.
|
Modifier and Type | Class and Description |
---|---|
class |
AbstractStreamingParser
Base class for streaming parsers.
|
class |
BitVectorLabelParser
Parser for parsing one BitVector per line, bits separated by whitespace.
|
class |
CategorialDataAsNumberVectorParser<V extends NumberVector>
A very simple parser for categorial data, which will then be encoded as
numbers.
|
class |
ClusteringVectorParser
Parser for simple clustering results in vector form, as written by
ClusteringVectorDumper . |
class |
LibSVMFormatParser<V extends SparseNumberVector>
Parser to read libSVM format files.
|
class |
NumberVectorLabelParser<V extends NumberVector>
Parser for a simple CSV type of format, with columns separated by the given
pattern (default: whitespace).
|
class |
SimplePolygonParser
Parser to load polygon data (2D and 3D only) from a simple format.
|
class |
SimpleTransactionParser
Simple parser for transactional data, such as market baskets.
|
class |
SparseNumberVectorLabelParser<V extends SparseNumberVector>
Parser for parsing one point per line, attributes separated by whitespace.
|
class |
TermFrequencyParser<V extends SparseNumberVector>
A parser to load term frequency data, which essentially are sparse vectors
with text keys.
|
Copyright © 2019 ELKI Development Team. License information.