test_calc_multifactor only uses a single factor

[humdings]

test_calc_multifactor and test_multifactor_beta only have a single factor passed to them. These tests should be made more robust. Anybody opposed to keeping a small csv of test data, or we can use random data with a fixed seed, thoughts?

[gusgordon]

It's a good idea, but I think these functions should get deprecated in the future in favor of having some flexible beta calculation function.

[humdings]

Sounds good, I think this PR is a step towards better regression methods but we will still need better tests either way.
#63