m

multiLayerPerceptrons

This package currently contains classes for training multilayer perceptrons with one hidden layer, where the number of hidden units is user specified. MLPClassifier can be used for classification problems and MLPRegressor is the corresponding class for numeric prediction tasks. The former has as many output units as there are classes, the latter only one output unit. Both minimise a penalised squared error with a quadratic penalty on the (non-bias) weights, i.e., they implement "weight decay", where this penalised error is averaged over all training instances. The size of the penalty can be determined by the user by modifying the "ridge" parameter to control overfitting. The sum of squared weights is multiplied by this parameter before added to the squared error. Both classes use BFGS optimisation by default to find parameters that correspond to a local minimum of the error function. but optionally conjugated gradient descent is available, which can be faster for problems with many parameters. Logistic functions are used as the activation functions for all units apart from the output unit in MLPRegressor, which employs the identity function. Input attributes are standardised to zero mean and unit variance. MLPRegressor also rescales the target attribute (i.e., "class") using standardisation. All network parameters are initialised with small normally distributed random values.
http://weka.sourceforge.net/doc.packages/multiLayerPerceptrons
GNU General Public License 3
University of Waikato, Hamilton, NZ
Eibe Frank
they using numeric rescales package zero ridge modifying over only multiplied bias problems random number available apart optionally many mlpregressor small optimisation minimum correspond before bfgs instances conjugated training identity faster normally latter standardisation classes overfitting specified classification gradient perceptrons minimise squared both distributed contains used where function values quadratic mlpclassifier parameters penalty control class multilayer size that input default employs output this decay weight attribute former currently from network penalised parameter corresponding mean implement find activation initialised standardised functions with error added layer averaged prediction descent logistic variance tasks units unit which weights hidden there target determined attributes local also user
Files download
File Operation
multiLayerPerceptrons-1.0.10.jar download
multiLayerPerceptrons-1.0.10.pom download
multiLayerPerceptrons-1.0.10-sources.jar download
Apache Maven 
<dependency>
  <groupId>nz.ac.waikato.cms.weka</groupId>
  <artifactId>multiLayerPerceptrons</artifactId>
  <version>1.0.10</version>
</dependency>
Gradle Groovy 
implementation 'nz.ac.waikato.cms.weka:multiLayerPerceptrons:1.0.10'
Gradle Kotlin 
implementation("nz.ac.waikato.cms.weka:multiLayerPerceptrons:1.0.10")
Scala SBT 
libraryDependencies += "nz.ac.waikato.cms.weka" % "multiLayerPerceptrons" % "1.0.10"
Groovy Grape 
@Grapes(
  @Grab(group='nz.ac.waikato.cms.weka', module='multiLayerPerceptrons', version='1.0.10')
)
Apache Ivy 
<dependency org="nz.ac.waikato.cms.weka" name="multiLayerPerceptrons" rev="1.0.10" />
Leiningen 
[nz.ac.waikato.cms.weka/multiLayerPerceptrons "1.0.10"]
Apache Buildr 
'nz.ac.waikato.cms.weka:multiLayerPerceptrons:jar:1.0.10'
Dependencies 
<parent>
  <groupId>org.sonatype.oss</groupId>
  <artifactId>oss-parent</artifactId>
  <version>7</version>
</parent>

compile

|-- nz.ac.waikato.cms.weka:weka-dev

test

|-- nz.ac.waikato.cms.weka:weka-dev

|-- junit:junit:3.8.2