This function test the null model. In this class, the test is either a F-test computed using stats::anova or a RESET test computed using lmtest::resettest.
Arguments
- dat
Data frame. A data frame containing all variables needed by the
null_formula.- null_formula
Formula. Formula for fitting the null model. Default is
null_formula = self$null_formula.- alt_formula
Formula. Formula for fitting the alternative model. Only used in F-test. Default is
alt_formula = self$alt_formula.- test
Character. A string representing the test. Either "F" or "RESET".
- power
Integer. A vector of integers representing the power of the variables that should be included. Only used in RESET test. Default is 2:3.
- power_type
Character. Argument passed to lmtest::resettest. "fitted", "regressor" or "princomp". Only used in RESET test. Default is "fitted".
Examples
# Instantiate
mod <- poly_model(shape = 2, sigma = 0.5)
mod
#>
#> ── <POLY_MODEL object>
#> y = 1 + x + include_z * z + e
#> - x: <RAND_UNIFORM object>
#> [a: -1, b: 1]
#> - z: <CLOSED_FORM object>
#> EXPR = (raw_z - min(raw_z))/max(raw_z - min(raw_z)) * 2 - 1
#> - raw_z: <CLOSED_FORM object>
#> EXPR = hermite(shape)((x - min(x))/max(x - min(x)) * 4 - 2)
#> - x: <RAND_UNIFORM object>
#> [a: -1, b: 1]
#> - e: <RAND_NORMAL object>
#> [mu: 0, sigma: 0.5]
#> Parameters:
#> - shape: 2
#> - include_z: TRUE
#> - sigma: 0.5
dat <- mod$gen(100000)
# F-test
mod$test(dat, test = "F")
#> $name
#> [1] "F-test"
#>
#> $statistic
#> [1] 147619.7
#>
#> $p_value
#> [1] 0
#>
# RESET test
mod$test(dat, test = "RESET", power = 2:4)
#> $name
#> [1] "RESET-test"
#>
#> $statistic
#> [1] 49206.78
#>
#> $p_value
#> [1] 0
#>