Implement Apriori Algorithm in WEKA

Launch WEKA

• Open WEKA → Click on Explorer

Load Your Dataset

• Go to the Preprocess tab
• Click Open File…
• Load a dataset in .arff or .csv format (preferably with nominal attributes only)
• Example: weather.nominal.arff (comes with WEKA)

Verify Attribute Types

• Ensure your attributes are nominal (not numeric)
• If needed, convert numeric to nominal:
  • Use Filter → unsupervised → attribute → NumericToNominal

Go to the Associate Tab

• Click the Associate tab
• Click Choose → Select weka.associations.Apriori

Configure the Apriori Algorithm (Optional)

• Click on Apriori to adjust parameters:
  • LowerBoundMinSupport: minimum support (e.g., 0.1 for 10%)
  • MinMetric: minimum confidence (e.g., 0.9 for 90%)
  • NumRules: number of rules to generate (e.g., 10)

Run the Algorithm

• Click Start
• WEKA will display the generated rules in the result output

Interpret the Output

Example:

Best rules found:

Outlook=Sunny ⇒ Play=No (conf: 0.8)

Humidity=High ⇒ Play=No (conf: 0.75)

Windy=False Play=Yes (conf: 0.6)

Each rule has support and confidence. These help you understand how frequent and reliable the rule is.

Save the Results (Optional)

Example of Association Rule Mining

Dataset (Transactions)

Step 1: Calculate Support

Support(X) = Fraction of transactions containing X

Rule:

Milk ⇒ Bread

Transactions with Milk and Bread together:

• T1: Milk, Bread ✓
• T2: Milk, Bread, Butter ✓
• T4: Milk, Bread, Butter ✓

text{Count(Milk & Bread)} = 3

• Total Transactions: 4

Support(Milk ⇒ Bread)=3/4=0.75=75%

Step 2: Calculate Confidence

Confidence(X⇒Y) = Probability of Y when X occurs

Confidence(Milk ⇒ Bread) = {Support(Milk & Bread)}/{Support(Milk)}

• Transactions with Milk:
  • T1 ✓
  • T2 ✓
  • T4 ✓

Support(Milk)=3/4=0.75

Confidence(Milk⇒Bread)=(3/4)/(3/4)=1.0=100%

Step 1: Find Frequent Itemsets

• {Milk, Bread} → Support = 3/4 = 75%
• {Bread, Butter} → Support = 3/4 = 75%

Step 2: Generate Rules

• Milk ⇒ Bread
  • Support = 75%
  • Confidence = 100%
• Bread ⇒ Butter
  • Support = 75%
  • Confidence = 75%