The Apriori algorithm смотреть онлайн

The slides are found at https://github.com/tommyod/Efficient-Apriori/blob/master/docs/presentation/apriori.pdf.

The Apriori algorithm uncovers hidden structures in data. The classical example is a database containing purchases from a supermarket. Every purchase has a number of items associated with it. We would like to uncover association rules such as (bread, eggs) implies (bacon) from the data. This is the goal of association rule learning, and the Apriori algorithm is arguably the most famous algorithm for this problem.

The Python implementation is found at https://github.com/tommyod/Efficient-Apriori, and the original paper by Agrawal et al, published in 1994, is found at https://www.macs.hw.ac.uk/~dwcorne/Teaching/agrawal94fast.pdf.

Contents
-------------
01:23 Motivating example - learning association rules
03:03 Support - the frequency of itemsets
04:33 Confidence - the conditional probability of a rule
06:03 Example of support and confidence
06:35 A naive algorithm
08:03 Overview of the Apriori algorithm
09:50 Generating itemsets via Apriori, example 1
11:15 Generating itemsets via Apriori, example 2
12:46 Pseudo-code for Phase 1 of the Apriori algorithm
14:16 Candidate generation and pruning
16:33 Checking if a set is a subset of another set
18:28 Sketch of Phase 2 of the Apriori algorithm
19:49 The Apriori algorithm on real data
21:47 Writing a Python implementation
25:25 Summary and references