Include BLOSUM and Grantham scores into pool of variables for running single + joint regressions.
Evaluate linear regression models for all possible combinations of variables including MPC + BLOSUM + Grantham and pick one with lowest AIC.
Minor changes:
Update documentation of additional_variables in function description run_regressions since it will now be added into combined variable pool for single + joint regressions.
Fix a bug causing single variables regression model with lowest AIC to output the model parameters instead of stats model object.
Replace all_model_aic with just the lowest AIC from each category of regression (single, additive, multiplicative and special) since having a long list of AICs from each combination is irrelevant in downstream steps.
Major Changes:
Minor changes:
additional_variables
in function descriptionrun_regressions
since it will now be added into combined variable pool for single + joint regressions.all_model_aic
with just the lowest AIC from each category of regression (single, additive, multiplicative and special) since having a long list of AICs from each combination is irrelevant in downstream steps.