Association Rules in Machine Learning
- 1. Akash A. Borse M.Tech (ACDS) PRN: 170847980005 22/03/2018
- 2. What is Machine Learning ? A major field from an umbrella term called “Artificial Intelligence”… Applying set of rules defined by humans over machines to make them self working and self learning.
- 3. What is Market –Basket Analysis ? A technique used by large retailers to identify associations among items. It works by looking for combinations of items that occurs frequently in transactions. It uses “Association Rules” for analyzing combinations of frequent items set.
- 4. Lets Understand Association Rule… It’s a technique to know how items are associated to each other. Factors to be considered to measure association: 1. Support 2. Confidence 3. Lift
- 5. 1. Support: • It states how popular an itemset is. It is measured by proportion of transactions in which itemset appears. 2. Confidence: • It states how likely an item Y is purchased when item X is purchased {X -> Y}. It is measured by the proportion of transaction with item X, In which item Y also appears 3. Lift: • It states how likely an item Y is purchased when item X is purchased, while controlling for how popular an item Y is.
- 6. Lets have an example… Source: kdnuggets