Creator of the dataset class
Syntax
data
data(d)
data(data,'attr1',val1,'attr2',val2,...)
data(data,symbols,vartypes,labels,variables)
data(classes,data,symbols,vartypes,labels,variables)
data(datafile,variables,symbols,vartypes,labels)
Description
data
creates an empty data object.
data(d)
clone the data object or fix a broken data object from its struct form.
data(data,'attr1',val1,'attr2',val2,...)
create a data object. data
are the data
(each piece of data is on a different line). The couples of attr1
,
val1
, ... are used to set the attributes.
data(data,symbols,vartypes,labels,variables)
create a data object. data
are the data
(each piece of data is on a different line). labels
is
a column vector of the labels (optional) corresponding to each piece of data.
variables
[cell], are the names of the columns of the data.
symbols
[cell] are the symbolic correspondance for the symbolic data.
vartypes
is a cell array containing the type of the variable for each
column ('continuous','discrete','symbolic','class'). There can only be one 'class'
variable at a time. If vartypes is not specified, the types are deduced from the
data
and the symbol field
data(classes,data,symbols,vartypes,labels,variables)
create a data object. data
are the data
(each piece of data is on a different line). labels
is
a column vector of the labels (optional) corresponding to each piece of data.
classes
contains the class associated with each piece of data.
variables
[cell], are the names of the columns of the data.
symbols
[cell] are the symbolic correspondance for the symbolic data.
data(datafile,variables,symbols,vartypes,labels)
create a data object from a mat datafile. datafile
is the
name of the mat file where the data are saved. The file must contain a series of
variables referenced by variables
[cell]. Each of this variable must be a
column vector([double] or [cell]) of length equal to the number of data points.
Methods
- addclass
add a class to the data set (obsolete)
- addfeature
Adds a new feature to the data set
- addpoint
Add a point to the dataset
- addsfeature
Adds a new symbolic feature to the data set
- addvariable
Add a new variable(feature) to the data set
- bootstrap
BOOTSTRAP creates n new datasets ready for Bagging purpose or cross validation
- btstrap2
BTSTRAP2 creates n new datasets ready for Bagging purpose or cross validation
- classgroup
Groups a few classes into more general classes
- classpart
selects a part (of the classes) of the data set
- csplit
Cut the dataset into n pieces for x validation purposes
- cut
Cut the dataset into n pieces for x validation purposes
- normalise
performs a denormalisation of the continuous data field of a dataset
- display
displays the dataset object
- double
converts the dataset object to double
- export
Export a data set to a comma delimited text file
- fs
FS performs a features selection on the dataset d
- fuzzyknn
performs fuzzy nearest neightbours algorithm
- fuzzyknn2
performs fuzzy nearest neightbours algorithm (matlab release)
- get
Access/query dataset property values
- getmiss
get missing values from a dataset
- gfkclus
performs Gustaffson Kessel clustering
- horzcat
Horizontal concatenation operator
- join
joins two data sets
- kmeans
performs kMeans clustering
-
- leave1out
LEAVE1OUT performs the leave-one-out validation and computes the confusion matrix,
error rates and Kappa inside a report structure
- mkmeans
performs multi prototypes clustering
- normalise
NORMALISE performs the linear normalisation of a learning dataset and transforms
the corresponding testing dataset if it is provided
- normstd
NORMSTD performs a normalisation of the continuous features of a learning set and a testing set
- pick
picks randomly chosen examples out of a data set
- placekernels
Place kernels to be used with smoothknn
- plot
plot a dataset
- postmining
compute the confusion matrix and error rates between supervision and prediction
- randfs1
RANDFS1 : a features subsets builder by sampling with replacement
- randfs2
RANDFS2 : a features subsets builder by sampling without replacement
- remPoint
remove a point from a data set
- set
Set object properties
- simple_knn
classifies a testing set using kNN
- simple_knn2
classifies the query using kNN (matlab release)
- simple_MLP
SIMPLE_MLP performs a one-hidden-layer neural network classification
- simpleknn
classifies the query using kNN
- simpleknn
classifies the query using kNN
- size
get the size of the data set
- smoothknn
classifies the query using smoothknn
- subsref
overload the subscript reference operator
- svm
Support Vectors Machine classification
- svminit
Initialise the Support Vectors for the svm classification
- vertcat
vertical concatenation operator
[ Help Desk ]