Association Rule Mining: Basic Theory & Practice

Name: Association Rule Mining: Basic Theory & Practice
Rating: 4.3 (52 reviews)

Theoretical Understanding and Market Basket Analysis with Python

Created byTakuma Kimura

Last updated 2/2022

English

What you'll learn

Basic theory of association rule mining
Basic metrics of association rule mining
Apriori algorithm
Market basket analysis with Python

Course content

5 sections • 21 lectures • 1h 42m total length

Introduction1:19
Explore the basics of association rule mining, its metrics and the apriori algorithm, with a practical Python hands-on section for applying data mining concepts in marketing and analytics.
What is Association Rule Mining?5:56
Explore what association rule mining is and how it differs from similar methods, revealing rules that show which products are purchased together to enable cross-selling and market basket analysis.
Comparison with Similar Methods4:13
Association rule mining discovers associations between items based on co-occurrence, distinguishing itself from clustering, which groups samples, and from collaborative filtering by focusing on item relationships in transactions.
Basic Terms3:22
Define basic terms in association rule mining—items, itemsets, antecedents, and consequents—using a market basket example, and explain that rules reveal co-occurrence, not causality, with support, confidence, and lift.
Basic Math: Set Theory2:35
Explore the basics of set theory, defining sets and elements and using epsilon to indicate membership. Demonstrate intersection and union of sets, using A and B as examples.
What is Association Rule Mining?

Support4:35
Explore how association rule mining uses metrics like support and confidence to measure itemset co-occurrence. Support equals the ratio of transactions containing focal itemset to all transactions.
Basic Math: Conditional Probability6:01
Explore conditional probability through dice-based examples, define P(A|B) as the probability of A given B, and compute it with P(A∩B)/P(B) and using frequencies.
Confidence7:26
Define confidence as probability of Y given X, the ratio of transactions with both X and Y to those with X; water to ham is 0.33, bread to ham 0.5.
Lift7:03
Define lift as the ratio of the conditional probability of Y given X to the probability of Y. Lift greater than 1 signals X and Y have an association.
Comprehension5:49
Explore how the basic association rule metrics—support, confidence, and lift—range from 0 to 1, 0 to 1, and 0 to infinity, and how confidence and lift derive from support values.
Summary1:29
Basic Metrics

The Role of Apriori Algorithm4:06
Explore the apriori algorithm to prune unimportant item sets in association rule mining, using metrics like support, confidence, and lift to manage exponential rule growth.
Apriori Principle4:52
Identify frequent itemsets using the Apriori principle, where all subsets of a frequent itemset are frequent, based on the anti-monotone property of support and a minsup threshold.
Apriori Algorithm: Phase 1ーFind Frequent Itemsets8:36
Apriori Algorithm: Phase 2―Association Rule Selection7:22
Summary0:57
Explore the apriori algorithm and how minsup and minconf prune rules, noting these thresholds are arbitrary and shape which rules are selected. Distinguish antecedent from consequent items to understand implications.
Apriori Algorithm

Dataset3:22
Dataset can be downloaded from the following link.
https://www.kaggle.com/roshansharma/market-basket-optimization/version/1
Data Preparation7:47
Prepare the dataset for association rule mining by converting transactions to lists, removing NaNs, and creating a one-hot encoded data frame using TransactionEncoder, enabling apriori rule analysis.
Basic Metrics7:02
Compute support, confidence, and lift from the onehot data using mean and logical_and, illustrating single and pair item supports with eggs and ground beef.
Apriori Algorithm7:57
Apply the Apriori algorithm in Python with mlxtend to mine frequent itemsets up to length four, prune by minsup, then generate and filter rules using confidence and lift.

Requirements

Knowledge of Python. Novice level is OK.

Description

Welcome to the association rule mining course. This course is an introductory course. You will learn basic knowledge of association rule mining in this course.

Association rule mining is a useful technique to explore associations between variables. It contributes to effective cross-selling and has been applied to construct recommender system in EC sites. We can use it not only in marketing analytics but also other fields in business analytics.

This course intends to provide you with theoretical knowledge as well as python coding. Theoretical knowledge is important to understand the algorithm of data mining, and it can be a useful foundation for more advanced learning.

This course consists of 4 sections. In the first section, you will learn what an association rule is. In Session 2, you will learn the basic metrics of association rule mining. Session 3 covers apriori algorithm that is a useful method to identify important associations between variables. Session 4 is a Hands-On chapter, where you will learn how to implement association rule mining in Python.

I’m looking forward to seeing you in this course!

Source of Pictures:

Course Image: Gerd Altmann from Pixabay
PV:
- Beer: Hans Braxmeie from Pixabay
- Pretzel: Couleur from Pixabay
- Potatoes: RitaE from Pixabay
- Diaper: PublicDomainPictures from Pixabay

Who this course is for:

Beginners in association rule mining

Association Rule Mining: Basic Theory & Practice

What you'll learn

Explore related topics

Course content

What is Association Rule Mining?5 lectures • 17min

Basic Metrics6 lectures • 32min

Apriori Algorithm5 lectures • 26min

Market Basket Analysis with Python4 lectures • 26min

Thank you!1 lecture • 1min

Requirements

Description

Who this course is for: