# Build a nuisance-configuration object for `mcee_general()` Creates a configuration list describing **how to obtain a nuisance function** used by `mcee_general`. You may either: 1. supply **known values** (bypasses learning), or 2. specify a **learning method** (e.g., GLM/GAM/RF/Ranger/SL) with a formula. ```r mcee_config_maker( target, method = NULL, formula = NULL, family = NULL, known = NULL, known_a1 = NULL, known_a0 = NULL, clipping = NULL, SL.library = NULL, ... ) ``` ## Arguments - `target`: Character; which nuisance to configure. One of `"p"`, `"q"`, `"eta"`, `"mu"`, `"nu"`. - `method`: Optional character learner name when *not* using known values. Supported: * `"glm"`, `"gam"`, `"lm"` (formula-based); * `"rf"` (randomForest), `"ranger"`; * `"sl"` (SuperLearner). Ignored if any of `known`, `known_a1`, `known_a0` is provided. - `formula`: RHS-only formula describing predictors for the learner (used when `method` is formula-based; ignored for `known`). For `method = "gam"`, `s()` terms are allowed (via `mgcv`). - `family`: Optional GLM/GAM family. If `NULL`, a default is chosen based on `target` and `method` (typically `binomial()` for `p`/`q`, `gaussian()` for `eta`/`mu`/`nu`). - `known`: Optional numeric scalar/vector of **known values** for the nuisance. Commonly used for `target = "p"` in MRTs when the randomization probability is known. - `known_a1, known_a0`: Optional numeric scalar/vector providing known values for the **treatment-specific** versions of a nuisance (e.g., `eta1`/`eta0`, `mu1`/`mu0`, `nu1`/`nu0`). If supplied, the function returns a "known" config and `method`/`formula`/`family` are ignored. - `clipping`: Optional numeric vector of length 2, `c(lower, upper)`, to truncate predictions (e.g., probabilities into $[\epsilon, 1-\epsilon]$). Validated but not required. - `SL.library`: Character vector of SuperLearner libraries (only used when `method = "sl"`). If `NULL`, defaults to `c("SL.mean", "SL.glm", "SL.gam")`. - `...`: Reserved for future extensions; currently ignored. ## Returns A named `list` describing the configuration. For known configs: ``` list( nuisance_parameter = , known = , known_a1 = , known_a0 = , clipping = ) ``` For learner configs: ``` list( nuisance_parameter = , method = , formula = , family = , clipping = , SL.library = ) ``` ## Details If any of `known`, `known_a1`, or `known_a0` is provided, the returned configuration is of type “known” and **no learner will be fit**. Otherwise, the configuration records the requested learner, formula, family, optional clipping, and (for SL) the library. Internally, helper validators ensure `method` is supported and `clipping` (if provided) is sane. Family defaults are chosen when `family = NULL` for GLM/GAM methods. ## Examples ```r # Known p (MRT randomization), GLM for other nuisances cfg_p <- mcee_config_maker("p", known = 0.5) cfg_q <- mcee_config_maker("q", method = "glm", formula = ~ dp + M) cfg_eta <- mcee_config_maker("eta", method = "glm", formula = ~dp) cfg_mu <- mcee_config_maker("mu", method = "glm", formula = ~ dp + M) cfg_nu <- mcee_config_maker("nu", method = "glm", formula = ~dp) # SuperLearner with default library (set explicitly if you prefer) # cfg_q_sl <- mcee_config_maker("q", method = "sl", formula = ~ dp + M, # SL.library = c("SL.mean","SL.glm","SL.ranger")) # Known treatment-specific outcome regressions (e.g., from external source) # cfg_eta_known <- mcee_config_maker("eta", known_a1 = rep(1, 100), # known_a0 = rep(0, 100)) ``` ## See Also `mcee_general`, helper constructors like `mcee_config_known()`, `mcee_config_glm()`, `mcee_config_gam()`, `mcee_config_lm()`, `mcee_config_rf()`, `mcee_config_ranger()`, `mcee_config_sl()`, `mcee_config_sl_user()`.

mcee_config_maker function