GpInsert: GpInsert

In jdeboer/gprediction: Google Prediction API client for R

Description

Train a Prediction API model.

Usage

  GpInsert(model_id, project, oauth, source_model = NULL,
    storage_location = NULL, storage_pmml_location = NULL,
    storage_pmml_model_location = NULL, model_type = NULL,
    training_instances = NULL, utility = NULL)

Arguments

model_id	The unique name for the predictive model.
project	The project associated with the model.
oauth	The httr OAuth2.0 token reference class object to use for authentication.
source_model	optional The Id of the model to be copied over.
storage_location	optional Google storage location of the training data file.
storage_pmml_location	optional Google storage location of the preprocessing pmml file.
storage_pmml_model_location	optional Google storage location of the pmml model file.
model_type	optional Type of predictive model (CLASSIFICATION or REGRESSION)
training_instances	optional List of outputs and csvInstances to train model on. An output is a string with the generic output value - could be regression or class label. A csvInstance is the input features for that instance.
utility	optional A class weighting function, which allows the importance weights for class labels to be specified (Categorical models only). A list of numeric keys (double)

Value

JSON response from the Google Prediction API converted into an R List.

References

Trainedmodels: insert https://developers.google.com/prediction/docs/reference/v1.6/trainedmodels/insert

jdeboer/gprediction documentation built on May 19, 2019, 8:24 a.m.