Getting more details about models trained with tabular data
In the Training a model in fastai with a curated tabular dataset recipe of Chapter 3, Training Models with Tabular Data, you trained a fastai model on a tabular dataset and used accuracy as the metric. In this recipe, you will learn how to get additional metrics for this model: precision and recall. Precision is the ratio of true positives divided by true positives plus false positives. Recall is the ratio of true positives divided by true positives plus false negatives.
These are useful metrics. For example, the model we are training in this recipe is predicting whether an individual's income is over 50,000. If it is critical to avoid false positives – that is, predicting an income over 50,000 when the individual has an income less than that – then we want precision to be as high as possible. This recipe will show you how to add these useful metrics to the training process for a fastai model.
