classify_qr: Binary Classifications from Quantile Regression
In Scott-Coggeshall/bigrq: Parallelized Quantile Regression with ADMM

classify_qr creates binary classifications from a quantile regression fit by comparing predicted values to a prespecified threshold.

1	classify_qr(newx, coef, y_transform, threshold)

`newx`	a design matrix containing the input values for predictions.
`coef`	a vector or matrix containing the estimated regression coefficient values. If a matrix (e.g. if you want classifications for model fits produced by a sequence of tuning parameter values), then the each column should correspond to a different fit.
`y_transform`	a vectorized function to be applied to the predictions.
`threshold`	the threshold to compare the predictions to.