More clarification on coef_all and coef_nonzero

Author: Qian, Hai

src/ports/postgres/modules/elastic_net/elastic_net.sql_in

@@ -293,7 +293,7 @@ elastic_net_predict(
 <DT>regress_family</DT>
 <DD>TEXT. The type of regression, either 'gaussian' ('linear') or 'binomal' ('logistic').</DD>
 <DT>coefficients</DT>
-<DD>DOUBLE PRECISION[]. Fitting coefficients.</DD>
+<DD>DOUBLE PRECISION[]. Fitting coefficients. If you use all the features in the following <em>ind_var</em>, then this should be <em>coef_all</em> in the fitting result table. If you use the selected features (features with non-zero coefficients) in <em>ind_var</em>, then this should be <em>coef_nonzero</em> </DD>
 <DT>intercept</DT>
 <DD>The intercept for the model.</DD>
 <DT>ind_var</DT>