Python: Keras

The R2 score explains the amount of variance in a model. The closer the score gets to 1, the more accurate the model is. The R2 score for Training set in this model was .774, which was average for this type of test. The R2 score for the Test model was .29, which was a low score. There was a significant difference between the two scores which may be a sign of overfitting, suggesting that the model is not generalizable for the test.
This model may have been more accurate if a slightly different test had been used. The Predicted R2 may have been a better option since it is better able to deal with the problem of overfitting. The major drawback of the Predicted R2 is that it is not as widely available as the standard R2, but since we were writing the code, availability would not have been and issue.
Another alternative would have been to simply report the data and allow the user to judge the value of that data for themselves.