Step 5: Compare performance of the models

Using the R2 values, we create a plot to compare the Minitab model and the Python model.
import matplotlib.pyplot as plt 
import seaborn as sns
ax = sns.regplot(x=pyPred,y=predictionsFromModelOps)
ax.set(xlabel='MLP Model Predictions', ylabel='MSS Model Predictions',title="Prediction Comparison between MSS and MLP Models")

R2List = {"MLP $R^2$":[R2Value_Python],
          "MSS $R^2$":[R2Value_MSS]}
R2Table = pandas.DataFrame.from_dict(R2List)
0 0.847415 0.862006

Let's also compare the MAD values.
MADList = {"MLP Mean Absolute Deviation":[MADValue_Python],
          "MSS Mean Absolute Deviation":[MADValue_MSS]}
MADTable = pandas.DataFrame.from_dict(MADList)
  MLP Mean Absolute Deviation MSS Mean Absolute Deviation
0 20131.055468 18561.12182

Based on these results, the models are comparable. Because the Minitab model is slightly better, can handle missing values, and is easier to monitor and interpret, we recommend Minitab Model Ops. This platform also allows for easy monitoring of many of your company's models.