- regularisation parameter for L2
- step parameter for optimizer
- max iterations for optimizer
- batch size, valid only for stohastic gradient descent
- which optimizer to use
- minimum change for loss function, used by optimizer
- include fitting of the intercept
- step parameter for optimizer
- include fitting of the intercept
- minimum change for loss function, used by optimizer
fit weights to the data provided by the implementation of the model
fit weights to the data provided by the implementation of the model
input DataSet with features and labels
initial weights, None if omitted
- regularisation parameter for L2
contains changes of the loss function value during optimisation at every step
contains changes of the loss function value during optimisation at every step
- max iterations for optimizer
- which optimizer to use
predict the probability of positive class for one vector of data
predict the probability of positive class for one vector of data
INDarray data vector
probability of positive class
predict class (1 or 0) for one vector of data
predict class (1 or 0) for one vector of data
INDarray data vector
predicted class
- batch size, valid only for stohastic gradient descent
contains weights after the fitting the model
contains weights after the fitting the model
SVM linear based on hinge loss. Creates a class of the SVM model