R/model.R

In dirmeier/netReg: Network-Regularized Regression Models

# netReg: network-regularized linear regression models.
#
# Copyright (C) 2015 - 2020 Simon Dirmeier
#
# This file is part of netReg.
#
# netReg is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# netReg is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with netReg. If not, see <http://www.gnu.org/licenses/>.


# netReg: network-regularized linear regression models.
#
# Copyright (C) 2015 - 2020 Simon Dirmeier
#
# This file is part of netReg.
#
# netReg is free software: you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation, either version 3 of the License, or
# (at your option) any later version.
#
# netReg is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with netReg. If not, see <http://www.gnu.org/licenses/>.


model <- function(p, q, family) {
  keras::keras_model_custom(function(self) {
    self$alpha <- init_vector(q)
    self$beta <- init_matrix(p, q)
    self$family <- family
    self$linkinv <- family$linkinv
    self$init_weights <- self$get_weights()

    self$reinit <- function() {
      self$set_weights(self$init_weights)
    }


    function(x, mask = NULL, training = FALSE) {
      eta <- linear.predictor(self$alpha, self$beta, x)
      if (self$family$family %in% c("gamma", "inverse.gaussian")) {
        eta <- tf$exp(eta)
      }
      self$linkinv(eta)
    }
  })
}