Lecture on A Priori Algorithm for Generating Strong Association Rules

Introduction

• Topic: Application of A Priori Algorithm.
• Goal: Generate strong association rules from a dataset.
• Minimum support: 40%.
• Minimum confidence: 70%.

Dataset Overview

• Consists of 5 transaction IDs.
• Products include bread, butter, milk, beer, cookies, diapers, etc.

Steps in Applying A Priori Algorithm

Step 1: Generate Frequent Item Sets

1. Generate Initial Item Sets:

   • Identify unique products: bread, butter, milk, diaper, beer.
   • Write these as 1-item sets.

2. Calculate Support Count:

   • Count product appearances (e.g., bread appears 3 times).
   • Determine support count threshold: 40% of 5 transactions = 2.
   • Identify frequent 1-item sets (e.g., all except cookies).

3. Generate 2-item Sets:

   • Create combinations of 2-item sets from frequent 1-item sets.
   • Calculate support for combinations (e.g., bread & butter: 3 times).
   • Identify frequent 2-item sets (e.g., bread & milk, bread & butter).

4. Generate 3-item Sets:

   • Identify combinations of 3-item sets.
   • Calculate support for each (e.g., bread, butter & milk: 2 times).
   • Identify frequent 3-item sets that meet support threshold.

5. Attempt to Generate 4-item Sets:

   • Not possible due to lack of enough items.

Step 2: Generate Association Rules

1. Calculate Confidence for Rules:

   • Formula: (Support of X & Y) / (Support of X).

2. Evaluate Rules Based on Confidence:

   • Example Rules: Bread & Butter
     • Bread → Butter: 100% confidence.
     • Butter → Bread: 100% confidence.
   • Example Rules: Bread & Milk
     • Bread → Milk: 67% confidence (not strong).
     • Milk → Bread: 100% confidence.
   • Repeat for other combinations like butter & milk, diaper & beer.

3. 3-item Set Rules Evaluation:

   • Bread, butter, milk combinations.
   • Use logic to write various possibilities.
   • Evaluate each for confidence and strength.

Conclusion

• The process involves generating frequent item sets and evaluating association rules.
• Strong rules are identified based on confidence exceeding the threshold.
• Practical application of A Priori algorithm.

Additional Information

• Encourage watching other solved examples for further understanding.
• Subscribe for more educational content.

This guide simplifies the steps necessary to apply the A Priori algorithm, illustrating how to identify frequent item sets and create strong association rules from a dataset.