NCA: Neighborhood Component Analysis (NCA).

Description Usage Arguments Value References

Description

A distance metric learning algorithm that tries to minimize kNN expected error.

Usage

1
2
3
4
NCA(num_dims = NULL, learning_rate = "adaptive", eta0 = 0.3,
  initial_transform = NULL, max_iter = 100, prec = 1e-08,
  tol = 1e-08, descent_method = "SGD", eta_thres = 1e-14,
  learn_inc = 1.01, learn_dec = 0.5)

Arguments

num_dims

Desired value for dimensionality reduction. If None, the dimension of transformed data will be the same as the original. Integer.

learning_rate

Type of learning rate update for gradient descent. Possible values are: - 'adaptive' : the learning rate will increase if the gradient step is succesful, else it will decrease. - 'constant' : the learning rate will be constant during all the gradient steps.

eta0

The initial value for learning rate.

initial_transform

If array or matrix that will represent the starting linear map for gradient descent, where d is the number of features, and d' is the dimension specified in num_dims. If None, euclidean distance will be used. If a string, the following values are allowed: - 'euclidean' : the euclidean distance. - 'scale' : a diagonal matrix that normalizes each attribute according to its range will be used.

max_iter

Maximum number of iterations of gradient descent. Integer.

prec

Precision stop criterion (gradient norm). Float.

tol

Tolerance stop criterion (difference between two iterations). Float.

descent_method

The descent method to use. Allowed values are: - 'SGD' : stochastic gradient descent. - 'BGD' : batch gradient descent.

eta_thres

A learning rate threshold stop criterion. Float.

learn_inc

Increase factor for learning rate. Ignored if learning_rate is not 'adaptive'. Float.

learn_dec

Decrease factor for learning rate. Ignored if learning_rate is not 'adaptive'. Float.

Value

The NCA transformer, structured as a named list.

References

Jacob Goldberger et al. “Neighbourhood components analysis”. In: Advances in neural information processing systems. 2005, pages 513-520.


jlsuarezdiaz/rDML documentation built on May 24, 2019, 12:35 a.m.