fit_true_model: Fit true regression model
In shiandy/bst235project:

Description Usage Arguments Details Value Functions

View source: R/fit_true_model.R

Fit true regression model E(Y_i | X_i) = alpha_0 + Phi(beta^T X_i), where Phi is the standard Normal CDF.

fit_true_model(xs, ys, init = NULL)

objective(alpha_betas, xs, ys)

gradient(alpha_betas, xs, ys)

`xs`	n x p matrix of predictors
`ys`	Length-n vector of outcomes
`init`	Initial guess for parameters
`alpha_betas`	Value of coefficients to evaluate the function. First entry must be the intercept, and the rest are the other coefficients.

Under the hood, the fitting is done using BFGS, as implemented in the optim function in R.

fit_true_model returns a length-(p + 1) vector of coefficients, where the first is the intercept and the rest are the beta values. If the fitting algorithm did not converge, fit_true_model will fail with an error.

objective and gradient return the value of the objective function and gradient evaluated for some coefficients alpha_beta, respectively.