Summary

You started your journey in this chapter with an introduction to classification models and their differences compared with regression models. You learned that the target variable for classifiers can only contain a limited number of possible values.

You then explored binary classification, wherein the response variable can only be from two possible values: 0 or 1. You uncovered the specificities for building a logistic regression model with TensorFlow using the sigmoid activation function and binary cross-entropy as the loss function, and you built your own binary classifier for predicting the winning team on the video game Dota 2.

After this, you went through the different performance metrics that can be used to assess the performance of classifier models. You practiced calculating accuracy, precision, recall, and F1 scores with TensorFlow, and also plotted a confusion matrix, which is a visual tool to see where the model made correct and incorrect predictions.

Then...