F = ((RSS0 - RSS1) / (p1 - p0)) / (RSS1 / (N - p1 - 1)) (ESL eq 3.13);
an intercept is always included.
Usage
morie_esl_f_test(model0, model1, X, y)
Arguments
- model0, model1
Column indices (1-based) of X in each model; model0
inside model1.
- X
Predictors without an intercept column.
- y
Response.
Value
Named list: statistic, p_value, df1, df2, rss0, rss1, p0, p1, n.
References
Hastie, Tibshirani & Friedman (2009), sec. 3.2.
Examples
i <- 1:20
X <- cbind(sin(i), cos(i), log(i))
morie_esl_f_test(1, 1:3, X, 1 + X[, 1] + 0.5 * X[, 3] + 0.1 * cos(7 * i))$statistic
#> [1] 277.816
