Association Rule Mining
Prerequisites to study: Unsupervised Learning
In Association Rule Mining we are interested in finding frequent patterns that occur in the input data and then we look at the conditional dependencies among these patterns.
Let suppose there are two events A and B which often occur together. Like, suppose we have a shop and have customers coming into the shop. So whenever customer A comes into the shop, customer B also comes to the shop. So if A is already in the shop, there is a very high chance that B is also in the shop. We look for these kinds of rules in Association Rule Mining which are conditional dependencies.
For example: In fault analysis, we would like to know which events occur more often along with the fault. Or we are interested in analyzing the transaction data where we are interested in knowing which of the items in a store are frequently bought together. For example, eggs, milk, and bread are bought together.
So we can find out frequently occurring patterns in the purchase data and then we can make rules out of those and find patterns.
A Transaction is a set of items bought together. A set or subset of items is called an item set in the Association Rule Mining Community.