BatchNormalization: applies the common batch normalization strategy on the activations of the parent layer.

SubsamplingLayer: subsample from groups of units of the parent layer by different strategies (average, maximum, etc.).DenseLayer: all units are connected to all units of its parent layer.ConvolutionLayer: applying convolution, useful for images and text embeddings.The following Neural Network Layers are available to build sophisticated architectures: FunctionalityĪll functionality of this package is accessible via the Weka GUI, the commandline and programmatically in Java.

The source code for this package is available on GitHub. The backend is provided by the Deeplearning4j Java library. It is developed to incorporate the modern techniques of deep learning into Weka. WekaDeeplearning4j is a deep learning package for the Weka workbench.

getOptions public 4j: Deep Learning using Weka.

Specified by: setOptions in interface OptionHandler Overrides: setOptions in class RandomizableClassifier Parameters: options - the options to parse Throws: - if parsing fails Version: $Revision: 10660 $ Author: Yasser EL-Manzalawy, FracPete (fracpete at waikato dot ac dot nz) See Also: LibSVMLoader, Generate probability estimates for classification Set the parameters C of class i to weight*C, for C-SVCĮ.g., for a 3-class problem, you could use '1 1 1' for equally Turns the shrinking heuristics off (default: on) Set tolerance of termination criterion (default: 0.001) Set cache memory size in MB (default: 40) Set the epsilon in loss function of epsilon-SVR (default: 0.1) WARNING: use only if your data has no missing values. WARNING: use only if your data is all numeric! Turns on normalization of input data (default: off) Set the parameter nu of nu-SVC, one-class SVM, and nu-SVR Set the parameter C of C-SVC, epsilon-SVR, and nu-SVR Set coef0 in kernel function (default: 0) Set gamma in kernel function (default: 1/k) Set degree in kernel function (default: 3) Note = ,ġ = polynomial: (gamma*u'*v + coef0)^degreeĢ = radial basis function: exp(-gamma*|u-v|^2) LibSVM classifier (e.g., confusion matrix,precision, recall, ROC score,Ĭhih-Chung Chang, Chih-Jen Lin (2001). LibSVM reports many useful statistics about LibSVM allows users to experiment with One-class SVM, Regressing SVM, and LibSVM runs faster than SMO since it uses LibSVM to build the SVM classifier. A wrapper class for the libsvm tools (the libsvmĬlasses, typically the jar file, need to be in the classpath to use this

Weka jar