Mining Multidimensional Association Rules
From Relational Databases and Data Warehouses
Lecture Based on Han & Kamber (2006), Section 6.4
🔹 Slide 1: Introduction
Association rule mining seeks relationships among data items.
Traditionally focused on single-dimensional transactional data.
Multidimensional association rules mine across attributes like
age, income, location, product.
Applicable to relational databases and data warehouses where
multiple attributes exist.
🔹 Slide 2: Key Definitions
Term Meaning
Multidimensional Involves ≥2 predicates (e.g., age,
Rule buys)
Inter-dimension Distinct predicates appear only
Rule once
Hybrid-dimension Repeated predicates (e.g., multiple
Rule buys)
Categorical Discrete values (e.g., color,
Attribute occupation)
Quantitative
Numeric values (e.g., age, income)
Attribute
age(X, "19–24") ∧ occupation(X, "student") → buys(X, "laptop")
💡 Example:
🔹 Slide 3: Three Approaches to Handling Quantitative Attributes
Approach Description
Predefined concept hierarchies
1️⃣ Static Discretization
used
2️⃣ Quantitative Association
Dynamic binning during mining
Rules
3️⃣ Distance-Based Association Clustering based on value
Rules proximity
🔹 Slide 4: 📊 Static Discretization of Quantitative Attributes
Replace numeric values with predefined ranges
Uses concept hierarchies (e.g., income: 20K–30K → “Low”)
Discretized attributes treated like categorical ones
Can use Apriori-like algorithms or data cubes
Diagram Suggestion:
Flowchart from Raw Data ➞ Concept Hierarchy ➞ Discretized ➞ Rule
Mining
🔹 Slide 5: Data Cubes for Efficient Mining
Store aggregate values for dimensions like age, income, buys
Cuboids represent combinations; base cuboid = full granularity
Efficient support/confidence computation
Diagram:
Multidimensional cube lattice
(age, income, buys) → slices & aggregations
🔹 Slide 6: 🔍 Quantitative Association Rules
Discretization occurs during mining
Focused on confidence/compactness optimization
age(X, "30–34") ∧ income(X, "42K–48K") → buys(X, "high-res TV")
Example:
Uses grid-based clustering, e.g., ARCS algorithm
Diagram:
2D Grid of Age vs Income → highlight clusters where rule applies
🔹 Slide 7: Rule Clustering Concept
Find close rules and merge to simplify
Cluster rules:
age=34, income=30K–40K → buys high-res TV
age=35, income=40K–50K → buys high-res TV
age=34–35 ∧ income=30K–50K → buys high-res TV
→ Combine into:
Diagram:
Grid with markers (X’s) merged into a rectangular region
🔹 Slide 8: 🧠 Distance-Based Association Rules
Use clustering algorithms to define intervals
Rules reflect approximate values (e.g., ~$200, not exact)
Replace support/confidence with:
Density Threshold
Degree of Association (distance-based)
Diagram:
Attribute values grouped by proximity (e.g., price clusters: [20,22],
[50,53])
🔹 Slide 9: Cluster-Based Rule Mining (Two-Phase)
Phase 1:
Apply clustering on attribute tuples
Use diameter/density metrics
Phase 2:
Mine rules of the form:
Cluster_A(age) → Cluster_B(income)
Diagram:
Clusters plotted with directed arrows showing association path
🔹 Slide 10: Summary
Multidimensional rules reveal rich, cross-attribute insights
Techniques vary by attribute type & data representation
Data cubes and concept hierarchies boost performance
Clustering empowers approximate and dynamic rule discovery
Would you like me to turn these diagrams into actual graphics for your
slides? Or take this content into Copilot Pages so we can refine speaker
notes and design together?