Skip to main content

Short note: Architectural aspects of DM

Short answer

A data mining system has data sources, a database/warehouse server, knowledge base, mining engine, pattern evaluation module, and user interface. Coupling with the database can be none, loose, semi-tight, or tight; tight coupling gives the best performance. Other aspects include centralised versus distributed mining and scalability.

Write a short note on Architectural Aspects of Data Mining.

The architecture of a data mining system describes how its components are organised and how it is coupled with a database or data warehouse.

Components

User interface
     |
Pattern evaluation module  <-- Knowledge base
     |                         (domain knowledge,
Data mining engine  ---------   thresholds)
     |
Database / data warehouse server
     |
Data cleaning, integration, selection
     |
Databases | Data warehouse | WWW | other sources
  1. Data sources: databases, data warehouse, web, files.
  2. Database or warehouse server: fetches the relevant data for the mining request.
  3. Knowledge base: domain knowledge, concept hierarchies, thresholds and user beliefs that guide the search.
  4. Data mining engine: functional modules such as classification, clustering, association and outlier analysis.
  5. Pattern evaluation module: filters interesting patterns by measures (support, confidence), interacting with the engine.
  6. User interface: lets the user specify tasks and view results and visualisations.

Coupling with a database or data warehouse

  • No coupling: the system reads from files and does not use any DB function (poor).
  • Loose coupling: uses some facilities of the DB, e.g. fetches data and stores results back.
  • Semi-tight coupling: some mining primitives (sorting, aggregation, statistics) are implemented in the database system.
  • Tight coupling: the mining system is fully integrated into the database or warehouse (best performance, preferred).

Other architectural aspects

Centralised vs distributed/parallel mining, scalability, client-server design, and integration with OLAP.

✓ Answer checked by Claude Opus 5.5 (high effort)

Related questions

All Data Mining chapter 7 questions →