Feature: add logistic regression and multiple regression as options

[Copilot]

Summary

This branch adds logistic regression and multiple regression as distinct analysis methods with automatic detection.

Changes

The changes made in this PR are:

1. Core functions: Created rwa_logit() for binary outcomes (logistic regression) and rwa_multiregress() for continuous outcomes (multiple regression)
2. Wrapper function: Converted rwa() to automatically detect method based on outcome variable (2 unique values → logistic, otherwise → multiple regression) with explicit method parameter override
3. API standardization: Unified output structure across both methods (consistent column naming, rescaled weights sum to 100, lambda matrices)
4. Numerical stability: Added safeguards in rwa_logit() for edge cases (clipping predictions away from 0/1, zero standard deviation handling)
5. Documentation: Added comprehensive roxygen2 docs with examples for both functions and new "Regression Methods" vignette
6. Plotting support: Extended plot_rwa() to handle both regression types with appropriate labeling
7. Test coverage: Added test suites for logistic regression, integration tests, and validation of numerical edge cases

Usage example:

# Automatic detection - binary outcome uses logistic regression
mtcars$high_mpg <- ifelse(mtcars$mpg > median(mtcars$mpg), 1, 0)
rwa(mtcars, "high_mpg", c("cyl", "disp", "hp", "wt"))

# Automatic detection - continuous outcome uses multiple regression  
rwa(mtcars, "mpg", c("cyl", "disp", "hp", "wt"))

# Explicit method selection
rwa(mtcars, "high_mpg", c("cyl", "disp", "hp", "wt"), method = "logistic")

---

Check

• roxygen2::roxygenise() has been run prior to merging to ensure that .Rd and NAMESPACE files are up to date.
• The R CMD checks pass (at least one or more of devtools::check() or rhub::check_for_cran())

Change Type

Please check the type of change your PR introduces:

• Bugfix
• Feature (incl. changes to visualizations)
• Technical design
• Build related changes (e.g. tests)
• Refactoring (no functional changes, no api changes)
• Code style update (formatting, renaming) or Documentation content changes
• Other (please describe):

---

(OPTIONAL) Note

Bootstrap confidence intervals currently only supported for multiple regression (not logistic regression).

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.