In this chapter, we saw different ways of analyzing the performance of supervised models. We started with regression model evaluation and then we moved on to classification models performance.
For regression models, we saw that the difference between predicted values and actual values is the most important measure. In this way, Rattle provides the Predicted versus Observed Plot.
We discovered that in classification, a false positive is different from a false negative. Based on this difference, we can create a confusion matrix and evaluate the performance of a classifier using different mechanisms such as a Risk Chart, or ROC Curve.
In this chapter, we've worked with Rattle because it provides the necessary tools to evaluate the performance ...