remove non-parametric fitting method

CLAUDE.md

@@ -4,7 +4,7 @@ This file provides guidance to Claude Code (claude.ai/code) when working with co
 
 ## Package Overview
 
-PKPDmap is an R package implementing Maximum A Posteriori (MAP) Bayesian estimation and non-parametric fitting for pharmacokinetic/pharmacodynamic (PK/PD) data. It depends heavily on [PKPDsim](https://github.com/InsightRX/PKPDsim) (also an InsightRX package) for ODE-based PK/PD simulation.
+PKPDmap is an R package implementing Maximum A Posteriori (MAP) Bayesian estimation for pharmacokinetic/pharmacodynamic (PK/PD) data. It depends heavily on [PKPDsim](https://github.com/InsightRX/PKPDsim) (also an InsightRX package) for ODE-based PK/PD simulation.
 
 ## Common Commands
 
@@ -39,8 +39,7 @@ The main entry point is `get_map_estimates()` (`R/get_map_estimates.R`), which s
 - `map` — standard MAP Bayesian with empirical Bayes (default)
 - `map_flat_prior` — MAP with flattened priors (reduced shrinkage)
 - `ls` — Least squares (nearly-flat MAP priors)
-- `np` — Non-parametric estimation on a user-supplied parameter matrix
-- `np_hybrid` — Hybrid MAP + non-parametric grid search
+
 
 ### Estimation Pipeline
 
@@ -52,9 +51,8 @@ get_map_estimates()
   │   ├── parse_omega_matrix()
   │   ├── parse_error()
   │   └── ...
-  ├── mle_wrapper()            # wraps optim() + Hessian calculation
-  │   └── ll_func_PKPDsim()   # likelihood: calls PKPDsim, applies error model + censoring
-  └── get_np_estimates()       # for np/np_hybrid methods
+  └── mle_wrapper()            # wraps optim() + Hessian calculation
+      └── ll_func_PKPDsim()   # likelihood: calls PKPDsim, applies error model + censoring
 ```
 
 ### Key Function Roles
@@ -66,7 +64,6 @@ get_map_estimates()
 | `R/ll_func_PKPDsim.R` | Computes log-likelihood by calling PKPDsim and applying residual error + censoring |
 | `R/calc_ofv_map.R` | Objective function value for MAP (includes prior penalty) |
 | `R/calc_ofv_ls.R` | OFV for least squares |
-| `R/get_np_estimates.R` | Non-parametric grid search fitting |
 | `R/parse_omega_matrix.R` | Converts various omega input formats to full covariance matrix |
 | `R/check_inputs.R` | Comprehensive upfront validation (throws descriptive errors) |
 | `R/run_sequential_map.R` | Batch MAP estimation across multiple individuals |

NAMESPACE

@@ -4,17 +4,14 @@ S3method(print,map_estimates)
 export(calc_ofv_ls)
 export(calc_ofv_map)
 export(create_block_from_cv)
-export(create_grid_around_parameters)
 export(create_iov_object)
 export(full_to_triangle)
 export(get_likelihood_of_data)
 export(get_map_estimates)
-export(get_np_estimates)
 export(join_blocks)
 export(ll_func_PKPDsim)
 export(lower_triangle_mat_size)
 export(map_shrinkage_control)
-export(np_settings_default)
 export(plot_eta)
 export(replace_list_elements)
 export(run_sequential_map)

R/get_map_estimates.R

@@ -8,8 +8,8 @@
 #' @param as_eta vector of parameters that are estimates as eta (e.g. IOV)
 #' @param weights vector of weights for error. Length of vector should be same 
 #' as length of observation vector. If NULL (default), all weights are equal. 
-#' Used in both MAP and NP methods. Note that `weights` argument will also 
-#' affect residuals (residuals will be scaled too).
+#' Note that `weights` argument will also affect residuals (residuals will be 
+#' scaled too).
 #' @param omega between subject variability, supplied as vector specifiying the
 #'  lower triangle of the covariance matrix of random effects
 #' @param weight_prior weighting of priors in relationship to observed data, 
@@ -49,10 +49,7 @@
 #' such as `Nelder-Mead` to avoid estimation failures.
 #' @param control list of options passed to `optim()` function
 #' @param allow_obs_before_dose allow observation before first dose?
-#' @param type estimation type, options are `map`, `ls`, and `np_hybrid`
-#' @param np_settings list with settings for non-parametric estimation (if 
-#' selected), containing any of the following: `error`, `grid_span`, grid_size`,
-#'  `grid_exponential`
+#' @param type estimation type, options are `map` and `ls`
 #' @param cols column names
 #' @param residuals show residuals? This requires an additional simulation so 
 #' will be slightly slower.
@@ -127,7 +124,6 @@ get_map_estimates <- function(
                       control = list(reltol = 1e-5),
                       allow_obs_before_dose = FALSE,
                       type = "map",
-                      np_settings = list(),
                       cols = list(x = "t", y = "y"),
                       residuals = TRUE,
                       verbose = FALSE,
@@ -354,23 +350,6 @@ get_map_estimates <- function(
     )
   )
 
-  #################################################
-  ## Non-parametric estimation (hybrid, based on
-  ## MAP fit as initial estimates)
-  #################################################
-  if(type == "np_hybrid") {
-    obj <- np_fit_wrapper(
-      obj = obj,
-      model = model,
-      regimen = regimen,
-      lagtime = lagtime,
-      data = data,
-      covariates = covariates,
-      weights = weights,
-      np_settings = np_settings
-    )
-  }
-
   #################################################
   ## Add g.o.f. info
   #################################################

README.md

@@ -1,6 +1,6 @@
 # PKPDmap
 
-MAP Bayesian and non-parametric data fitting for PK(PD) models.
+MAP Bayesian fitting for PK(PD) models.
 
 ## Description
 
@@ -9,8 +9,6 @@ This R library implements various methods for fitting of individual PK(PD) data.
 - `map`: MAP Bayesian: standard maximum a priori estimation of empirical Bayes estimates
 - `map_flat_prior`: MAP Bayesian with flat prior: reduced weight of prior distribution, i.e. will put more trust in the data and hence reduce shrinkage (user can control the weighting).
 - `ls`: Least Squares: not *truly* least squares but rather special case of MAP Bayesian with flat priors, i.e. with (almost) fully flat prior.
-- `np`: non-parametric estimation. Performs non-parametric estimation based on user-specified parameter matrix of previously studied subjects.
-- `np_hybrid`: hybrid MAP and non-parametric extended grid. Will first perform standard MAP estimation (potentially with flattened prior), and subsequently perform non-parametric estimation on grid of support points around MAP estimates.
 
 ## Installation