I'd prefer to make smaller logical steps that I can understand. You can estimate stiffness (E) from SPT N-Value if you know the soil type (see tables in J.Bowles). Then then you can estimate the modulus of subgrade reaction (K) from the stiffness. Then finally you can estimate CBR from the dimensions of the CBR.
You can use Winterkorn & Fangs foundation engineering book (section 16.8) to convert stiffness to subgrade reaction.
By making these kinds of smaller steps you get a sense of how sketchy some of the correlations are. Keep in mind that CBR and SPT are both index tests, they're don't measure engineering properties of soils. I prefer using correlations between standard index tests and engineering properties (strength, stiffness).