Summary
In this chapter, we learned how to apply quantum annealing to build a strong classifier from several weak ones. We started with the general principles of quantum boosting and its corresponding QUBO formulation.
We then illustrated the application of the QBoost algorithm to solving a practical real-world financial problem, namely predicting credit card clients defaulting on their payments. The chosen dataset is reasonably large and complex enough to provide a meaningful challenge while remaining easy to understand and interpret the obtained results.
It is important to have an objective comparison with the corresponding classical counterparts. With this in mind, we introduced several classical classifiers based on the concepts of a feedforward neural network and a decision tree. We benchmarked QBoost against the MLP and gradient boosting models using such metrics as accuracy, precision, and recall.
In the next chapter, we will learn how quantum annealing can assist in training powerful...
