In many practical applications data comes in the form of lists (ordered or unordered): grocery lists, playlists, visited locations or URLs, app logs, and so on. Sometimes those lists are generated as a byproduct of business processes, but they still contain potentially useful information and insights for process improvement. To extract some of that hidden knowledge, one can use a special kind of unsupervised learning algorithm—association rule mining. In this chapter, we are going to build an app that can analyze your shopping lists to find out your preferences in the form of rules such as "If you've bought oatmeal and cornflakes, you also want to buy milk." This can be used to create an adaptable user experience, for instance, contextual suggestions or reminders.
In this chapter, we will cover the following topics:
- Association rules
- Association measures
- Association rule mining algorithms
- Building an adaptable user experience