Matrix files created by REGDIAG
The following matrices are available for additional computations:
REG.M vector of regression coefficients
REGS.M regression coeffients and their standard errors
REG_CORR.M correlations of regression coefficients
RG.M vector of miscellaneous scalars (see next page)
The covariance matrix of regression coefficients can be computed by
/COV REG_COV,REG_CORR.M,REGS.M
The vector RG.M has the following contents (when displayed by MAT LOAD).
REGDIAG statistics from data DECA
RG.M
/// REGDIAG
n 48.000 / # of cases
k 6.000 / # of regression coefficients
const 1.000 / 1 if constant term included, 0 otherwise
df 42.000 / n-k
Yvar 26132.000 / variance of regressand
SST 1228203.979 / total sum of squares
SSE 445535.121 / residual sum of squares
SSR 782668.858 / SST-SSE
MSE 10607.979 / mean square error
Resvar 10607.979 / MSE
R 0.798 / multiple correlation coefficient
R2 0.637 / R^2
kappa 93.968 / condition number
DW 1.246 / Durbin-Watson statistics
F 14.756 / Overall F-test: SSR/(k-1)/MSE (only when const is 1)
Pr_F 1.000 / P-value of F-test