Package pylearn :: Package algorithms :: Module kernel_regression
[hide private]

Module kernel_regression

source code

Implementation of kernel regression:

Classes [hide private]
Implementation of kernel regression: * the data are n (x_t,y_t) pairs and we want to estimate E[y|x] * the predictor computes f(x) = b + \sum_{t=1}^n lpha_t K(x,x_t) with free parameters b and alpha, training inputs x_t, and kernel function K (gaussian by default).
A kernel predictor has parameters theta (a bias vector and a weight matrix alpha) it can use to make a non-linear prediction (according to the KernelPredictorEquations).
Functions [hide private]
kernel_predictor(inputs, params, *otherargs) source code
Variables [hide private]
  row_vector = theano.tensor.DimShuffle((False,), ['x', 0])
  col_vector = theano.tensor.DimShuffle((False,), [0, 'x'])

Imports: OfflineLearningAlgorithm, T, prepend_1_to_each_row, as_scalar, AutoName, theano, numpy