Member-only story
A Guide to Association Rule Mining
Create insights from frequent patterns using market basket analysis with Python
Association rule mining is a rule-based machine learning technique used to find frequent patterns in a data set. Frequent patterns may include frequent itemsets that are usually bought together or subsequences that are bought in sequence. For example, cookies and coffee can be frequent itemset for a cafe, and a laptop and external monitor can be a subsequence for an electronics store. Finding frequent patterns in a transactional database and detecting associations between items is an extremely popular data science use case. Some application areas are item recommendations, cross-selling, promotion design, customer behavior analysis, and inventory management.
In this article, we will cover market basket analysis technique for association rule mining and will answer questions including but not limited to:
- if a customer buys item A how likely will she also buy item B? Is there a positive or negative correlation or it happens totally at random?
- which items should be placed next to each other in a store or content page of an app?
- which items can be bundled together in a promotional activity or recommended after showing interest in a certain item?
