Apache Spark Deep Learning Cookbook

How it works...

This section explains how the model performance is evaluated.

1. In order to compute the accuracy of our model, it is important to be able to identify how accurate our predictions were. Often, this is best visualized using a confusion matrix cross table that shows correct and incorrect prediction scores. We create a confusion matrix using the crosstab() function off the df_predicted dataframe that shows us we have 964,980 true negative predictions for labels that are 0 and we have 48,034 true positive predictions for labels that are 1, as seen in the following screenshot:

2. We know from earlier in this section that there are ...