Apriori Algorithm is a concept that is used in data mining in the digital marketing world. This is an algorithm introduced in 1994 by R. Agarwal and R. Srikant. Apriori algorithm is very important in the data mining process.
There are three significant components in this algorithm
To understand the concept better, here is an example.
Suppose you go to the supermarket to buy bread, there are a lot of chances you would also pick up jam and butter along with it. This is because people eat bread with butter or jam or both. This customer study is researched well by the owners of the supermarket. Now, what he did is create a bundle containing all these three items together for the following reasons.
- For making it easier for the public to pick all things at one place without searching a lot
- To allow discounts because more items are included and he can give discounts
- The discounts and the bundled items will increase his sales
- Both the user and the seller is benefited
Let us explain these three concepts with this example so that apriori algorithm is clear.
It is basically the default popularity of any item. You can calculate support like this
Support (jam) = (Transactions involving jam) / (Total transactions)
In this example, confidence is surety that the customers will buy both the butter and jam. You can calculate confidence like this
Confidence = (transactions involving both bread and jam) / (total transactions involving jam)
With this, you will find out how many customers bought both the butter and jam together.
In this same example, lift is the ratio of the sale of bread while you sale jam. You can calculate lift like this
Lift = (Confidence (jam – bread) / (Support (jam))
It shows what the likelihood is of the customers buying both jam and bread together than before. If the Lift value is greater than 1 then the customers are likely to buy that combination and if the lift value is less than 1 then the customers are not likely to buy that combination.