Fits every available curve to the same points and reports how well each follows the measured
lactate, to help choose fit in lactate_threshold(). The statistics are computed on the measured
stages (plus the resting sample when include_baseline = TRUE).
Arguments
- x
The result of
lactate_threshold().
Value
a tibble with one row per curve: fit, used (the curve chosen in lactate_threshold()),
n_points, parameters, r_squared, rmse and see (standard error of the estimate) in mmol/L, and
max_abs_residual and the leave-one-out prediction_error (root-mean-square) in mmol/L.
Details
A curve with more parameters always follows the points more closely, so the RMSE favors it. The prediction error does not: each stage is predicted by the curve fitted without it, which punishes a curve that bends to follow single points (leave-one-out cross-validation).
Examples
results <- lactate_threshold(demo_data, intensity, lactate, plot = FALSE)
lactate_fit_quality(results)
#> # A tibble: 4 × 9
#> fit used n_points parameters r_squared rmse see max_abs_residual
#> <chr> <lgl> <int> <int> <dbl> <dbl> <dbl> <dbl>
#> 1 3rd degree… TRUE 7 4 1.000 0.0488 0.0746 0.0758
#> 2 4th degree… FALSE 7 5 1.000 0.0433 0.0810 0.0703
#> 3 B-spline FALSE 7 5 1.000 0.0223 0.0417 0.0364
#> 4 Exponential FALSE 7 3 1.000 0.0563 0.0745 0.100
#> # ℹ 1 more variable: prediction_error <dbl>