Table 1.

Values of $\beta$ for the data generation process in scenario 1 and scenario 2

	${\mathit{\beta }}_{\mathbf{0}}$	${\mathit{\beta }}_{\mathbf{1}}$	${\mathit{\beta }}_{\mathbf{2}}$	${\mathit{\beta }}_{\mathbf{3}}$	${\mathit{p}}^{\mathbf{\prime }}$	n
Scenario 1	0.00	3.00	−2.00	0.50	22	2000
Scenario 2
Unbalanced/strong	−2.65	3.60	−2.20	−1.00	197	100
Balanced/strong	0.00	3.60	−4.00	−1.00	197	100
Unbalanced/weak	−2.75	0.60	−2.50	−1.00	197	100
Balanced/weak	0.50	0.60	−2.50	−1.00	197	100

Values of $\beta$ associated with non-informative features are not reported since they were all null. Last two columns are $p^{\prime }$ number of non-informative features and n number of observations