Introduction

We will address four types of problems that involve different kinds of decision making.

• Designing a live recommendation engine: Recommendation engines determine which items some users would most likely want to buy or visit a particular website for, given that we have some knowledge about their previous actions, as well as knowledge of all user behavior that was observed on an e-commerce website. We'll design such an engine based on the measurement of co-occurrence of events and lay down a data structure, making it possible for us to effectively process all user data as it flows through the system.

• Resolving cost and profit optimization problems: In this recipe, we will use an interesting technique that is widely used in artificial intelligence—branch and bound—in order to resolve a special case of optimization problems, where solutions can only be expressed as a linear combination of integer values. We'll apply this technique in the particular case of a bakery. We will decide...