QDA_eval: graphical model model evaluation using QDA as a classifier

In JointNets: End-to-End Sparse Gaussian Graphical Model Simulation, Estimation, Visualization, Evaluation and Application

Description

graphical model model evaluation using QDA as a classifier

Usage

QDA_eval(train, valid, test, lambda_range, v_seeking_length = 10,
  method = "diffee", ...)

Arguments

train	a list of training data
valid	a list of validation data
test	a list of test data
lambda_range	a vector of lambda values to train to given method, eg c(0.1,0.2,0.3)
v_seeking_length	second hyperparameter length, default to 10
method	name of the method to be evaluated
...	optional parameters passed to your method from JointNets package

Value

covriance matrix / kendall tau correlation matrix

Examples

library(JointNets)
data("nip_37_data")
split = train_valid_test_split(nip_37_data,c(0.8,0.1,0.1),10000)
train = split[["train"]]
valid = split[["valid"]]
test = split[["test"]]
v_seeking_length = 2
lambda_range = seq(0.5,1, length.out = 2)
result = QDA_eval(train,valid,test,lambda_range, v_seeking_length, method = "diffee")
result[["best test accuracy"]]

JointNets documentation built on July 30, 2019, 1:02 a.m.