Self-Organizing Molecular Field Analysis (SOMFA) comes with a built-in regression methodology, the Self-Organizing Regression (SOR), instead of relying on external methods such as PLS. In our recent paper, we presented a proof of the equivalence among SOR, SIMPLS, and NIPALS with one principal component. Thus, the modest performance of SOMFA on complex datasets can be primarily attributed to the low performance of the SOMFA regression methodology. In this paper, the SOR, originally proposed along with the SOMFA analysis, and external multi-component regression methods Multi-Component SOR (MCSOR) and SIMPLS are used to evaluate the performance of SOMFA. The performance gain achieved by external regression tools is assessed using the TBG and SADLER benchmark datasets and a large and diverse xenoestrogen dataset containing activity data for five different estrogen receptors. The effect of polarizibility descriptor and two superposition techniques on the predictive ability of SOMFA is also evaluated. The results clearly indicate that for diverse datasets SOMFA clearly benefits from the use of external regression tools instead of the SOR regression. On the other hand, no clear difference was observed between the two superposition techniques. The polarizability descriptor generated predictive models as a stand-alone descriptor but clear improvement in the accuracy of the prediction is achieved when the polarizability descriptor is combined with the electrostatic field descriptor.