---
title: "Workflow Guide"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Workflow Guide}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r, include = FALSE}
knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)
```

`SelectBoost.FDA` now covers several distinct FDA workflows: raw-grid designs,
basis and FPCA preprocessing, grouped stability selection, FDA-aware
`SelectBoost`, perturbation surfaces, diagnostics, and benchmarking. This
vignette is the entry point for the rest of the documentation.

## Start here

If you are new to the package, use the vignettes in this order.

1. [Discretized Curves and Grouped Stability Selection](discretized-curves.html)
   starts from raw functional curves and builds the main `fda_design` object.
2. [Basis and FPCA Workflows](basis-fpca-workflows.html) shows how to fit and
   reuse spline-basis or FPCA preprocessing.
3. [SelectBoost for Dense Spectra](selectboost-spectra.html) focuses on the
   dense, highly correlated setting where FDA-aware `SelectBoost` is most
   useful.
4. [Methods, Calibration, and Formula Workflows](methods-and-formulas.html)
   covers the broader modeling API: selector backends, formulas, and
   calibration helpers.

## Diagnostics and Validation

Once the core workflow is clear, the next vignettes explain how to inspect and
compare results.

1. [Association Diagnostics for Functional Predictors](association-diagnostics.html)
   explains correlation, neighborhood, hybrid, and interval association
   structures.
2. [Perturbation Grids for FDA Selection](perturbation-grid.html) introduces
   the two-parameter selection surface indexed by `q` and `c0`.
3. [Monotonicity and Precision-Recall Paths](monotonicity-and-pr-paths.html)
   shows how to summarize those surfaces without relying on a plotting
   backend.

## Benchmark Material

The last two vignettes are benchmark-oriented.

1. [Simulation and Benchmark Workflows](simulation-and-benchmarks.html)
   covers the general simulation engine and method-comparison utilities.
2. [Focused Benchmark Workflow](focused-benchmark-workflow.html) focuses on the
   saved benchmark tables and the dedicated benchmark driver.

## Minimal Orientation Example

The package entry point is still the FDA-native design object.

```{r}
library(SelectBoost.FDA)

sim <- simulate_fda_scenario(
  n = 18,
  grid_length = 10,
  include_scalar = FALSE,
  seed = 1
)

sim$design
head(selection_map(sim$design))
```

From that point, the package branches into grouped stability selection,
FDA-aware `SelectBoost`, or benchmark workflows depending on the question you
want to answer.