Function ols_granger_single_dim

fn ols_granger_single_dim(y: &[f64], x: &[f64], lag: usize) -> (f64, f64, usize)

Fit OLS for a single dimension and compute residual sum of squares.

Restricted model: y_d(t) = Σ_l β_l · y_d(t-l) + ε Unrestricted model: y_d(t) = Σ_l β_l · y_d(t-l) + Σ_l γ_l · x_d(t-l) + ε

Returns (rss_restricted, rss_unrestricted, n_obs).