phi_est: Estimate Phi on half of data
In jackiemauro/CACE:

Take in the estimates done of the nuisance function on the training data. Get predictions using these on the test data. Input those into the formula for your real parameter. Output the estimate of your parameter for this train/test combo.

1
2
3

phi_est(y, a, z, cov, ymean, amean, p, delta = 20)

phi_est(y, a, z, cov, ymean, amean, p, delta = 20)

`y`	is the outcome vector
`a`	is the treatment vector
`z`	is the instrument vector
`cov`	is a dataframe of covariates
`ymean`	is the mean estimates for the outcome variable from the mean_est function
`amean`	is the mean estimates for the treatment variable from the mean_est function
`p`	the conditional density estimation object from the propscore_est funciton
`delta`	the level you want to shift by
`dat`	is a dataset that is being used as your training dataset overall
`m`	a list of your mean estimates from the mean_est function
`p`	the conditional density estimation object from the propscore_est funciton