Input parameter | Explanation |
SEED | Random number seed (should be a positive integer) |
ALEVEL | Significance level of the statistical test (Type I error) |
P | Event proportion (response probability) |
NITER | Number of iterations performed |
N_REPEAT | Number of iterations performed |
| ^{*}NITER and N_REPEAT should be the same number |
PATH | Directory in which results are saved |
TABLE | Table name for saved results |
R | Number of categorized groups |
| Example: continuous = 1, median = 2, tertile = 3, quantile = 4 |
| ^{*}If model includes nominal variables, R should be >1 |
CHANGE_POINT | Change point (see Figure 1) |
Regression coefficients for the covariates in the full model, except for predictors and intercept, specified as: | |
MODEL_1 | %NRSTR(α_{1}^{*}X1+, , + β_{i}X_{i}) |
MODEL_2 | %NRSTR(α_{1}^{*}(the value of change_point) + α_{2}^{*}(X1 − (the value of change_point) +, , + β_{i}X_{i}) |
| α and β are the given regression coefficient values |
^{*}If model is linear, the regression coefficients α_{1} and α_{2} are the same | |
Sample size, mean, standard deviation, skewness, kurtosis, and correlation are specified as: | |
Example DATA a (type=CORR); LENGTH _TYPE_ $40; INPUT _NAME_ $_TYPE_$ X1 X2 ; IF TRIM(LEFT(_TYPE_))=’N’ THEN call symput(‘NSP’, X1); CARDS; . MEAN 70 50 . STD 4 5 . N 300 300 X1 CORR 1 0 X2 CORR 0 1 ; RUN; | |
^{*}If only one covariate is defined, the correlation should be set to 1. The sample size of all covariates should be the same. | |
SKW_KRT | %NRSTR ({skewness 1 kurtosis 1, skewness 2 kurtosis 2, }) |
^{*}If covariates are normally distributed, both skewness and kurtosis are set to 0. | |
LIST_VARNAME | %NRSTR (X1, X2, , Xi); list of variable names in A of above dataset |
MIN | Minimum value of a continuous variable |
MAX | Maximum value of a continuous variable |
SUB_GROUP | Number of subgroups |
CATEGORIZATION | %NRSTR (list of covariates to be categorized) |
CATEGORIZATION_R | %NRSTR (list of new covariate names after categorization) |
CONTI_MODEL | %NRSTR (list of covariates in a continuous logistic regression model) |
^{*}Even if some parameters are not needed, please assign all parameters and specify necessary variables in a logistic regression model. |