Skip to contents

Linear Model for morphometric data

Source: R/stat_lm.R
stat_lm.Rd

Fit linear models with coefficient data as response or predictor.

Usage

stat_lm(data, formula, ...)

Arguments

data

A tibble with coefficient and/or scalar columns

formula

A formula specifying the model. Can be:

  • Shape regression: coe ~ length (shape depends on predictors)

  • Response regression: length ~ coe (response depends on shape)

  • Multivariate response: cbind(length, width) ~ coe

  • Multiple predictors: coe ~ length + temperature

  • Interactions: coe ~ length * temperature

...

Additional arguments passed to stats::lm()

Value

An object of class c("stat_lm", "momstats") containing:

  • data: Original tibble (unchanged)

  • model: The stats::lm() object

  • method: "lm"

  • direction: "shape_regression" or "response_regression"

  • call: The function call

  • formula: Formula used

  • response_col: Response column name(s)

  • predictor_cols: Predictor column name(s)

  • r_squared: R^2 value(s)

  • adj_r_squared: Adjusted R^2 value(s)

Details

stat_lm() provides a unified interface for linear modeling with morphometric data.

Formula directions

The function automatically detects model direction from the formula:

Shape regression (coe ~ predictors):

  • Fits multivariate regression with all harmonics as response

  • Proper treatment of shape as correlated multivariate outcome

  • Tests whether predictors affect shape

  • Example: boteft %>% stat_lm(coe ~ length)

Response regression (response ~ coe):

  • Fits regression with unfolded harmonics as predictors

  • Tests whether shape predicts a response variable

  • Example: boteft %>% stat_lm(length ~ coe)

Multivariate response (cbind(y1, y2) ~ coe):

  • Fits multivariate regression with multiple responses

  • Example: boteft %>% stat_lm(cbind(length, width) ~ coe)

Interactions and multiple predictors

Standard R formula syntax is supported:

  • coe ~ length + temperature: Additive effects

  • coe ~ length * temperature: With interaction

  • coe ~ poly(length, 2): Polynomial terms

Getting results

Use collect() to add predictions to your data:

lm_fit <- boteft %>% stat_lm(coe ~ length)
boteft_pred <- collect(lm_fit)  # Adds fitted values

Use transduce() to predict at new positions:

new_shapes <- transduce(lm_fit, tibble::tibble(length = seq(50, 150, by = 10)))

See also

stats::lm(), collect.stat_lm(), plot.stat_lm(), transduce.stat_lm()

Examples

if (FALSE) { # \dontrun{
# Shape regression
lm1 <- boteft %>% stat_lm(coe ~ length)
lm2 <- boteft %>% stat_lm(coe ~ length + type)
lm3 <- boteft %>% stat_lm(coe ~ length * type)

# Response regression
lm4 <- boteft %>% stat_lm(length ~ coe)

# Multivariate response
lm5 <- boteft %>% stat_lm(cbind(length, width) ~ coe)

# Plot results
plot(lm1)
plot(lm1, type = "residuals")

# Get predictions
boteft_pred <- collect(lm1)

# Transduce to new values
new_shapes <- transduce(lm1, tibble::tibble(length = seq(50, 150, by = 10)))
} # }