Home > Topics > Data Mining and Business Intelligence > Kinds of Mining Data
← PreviousNext →

Kinds of Mining Data 📂🌐

Data mining isn't restricted to simple Excel sheets. In the modern world, data comes in many "flavors," each requiring different mining techniques.

Loading stats…

1. Relational Databases

These are the standard tables (Rows and Columns) found in most company software like ERPs.

  • Highly Structured: Data is organized into tables with predefined relationships (Keys).
  • Scalability: Mining must handle databases that grow by millions of rows daily.
  • SQL Based: Initial data selection is done using standard SQL queries before mining begins.
  • Integrity: Relational data is usually of higher quality because of strict database rules.

2. Transactional Data

This data records high-speed "events" rather than static information.

  • Time-Sensitive: Every transaction has a timestamp (e.g., Exactly 10:45 AM).
  • ACID Properties: Transactional data is reliable because it follows strict database safety rules (Atomicity, Consistency, Isolation, Durability).
  • Fast Growth: A single bank can generate billions of transaction rows every month, requiring high-performance mining.
  • Association Mining: This is the primary home for "Market Basket Analysis."

3. Data Warehouses

A central repository where data from multiple systems is integrated and summarized.

  • Multidimensional: Data is stored in "Cubes" (e.g., Time vs. Region vs. Product).
  • Cleaned & Formatted: Unlike raw databases, warehouses contain data that has already been standardized.
  • Historical Focus: While databases focus on "What is happening now," warehouses focus on "What has happened over the last 10 years."

4. Spatial and Temporal Data

This data involves the dimensions of "Where" and "When."

  • Spatial (Geographical): Mapping data, GPS coordinates, and satellite imagery. Used to find the best location for a new store.
  • Temporal (Time-Series): Data that tracks a single variable over time (like stock prices or humidity).
  • Sequence Data: Analyzing the order of events (e.g., a customer buys a camera, THEN a tripod, THEN a memory card).

5. Multimedia Data

Mining content that is not made of text or numbers.

  • Feature Extraction: The computer identifies "shapes" in images or "rhythms" in music to find patterns.
  • Image Mining: Used in medical fields to identify disease patterns in X-rays or MRI scans.
  • Video Mining: Security systems using facial recognition or traffic cameras detecting accidents in real-time.
  • Audio Mining: Analyzing voice recordings to detect the "Sentiment" (Happy or Angry) of a customer in a call center.

6. Stream and Web Data

Data that is constantly "flowing" and never stops.

  • Stream Data Mining: Analyzing data that arrives at high frequency (e.g., sensor data from an airplane engine).
  • World Wide Web (Web Mining): Mining the structure of websites (web links), the content (web pages), and the usage (web logs).

Summary

  • Relational: Traditional table-based mining.
  • Transactional: Analyzing high-speed business events.
  • Spatial: Mining map-based and location data.
  • Temporal: Finding patterns in time-series data.
  • Multimedia: Extracting meaning from images and videos.

Quiz Time! 🎯

Loading quiz…

← PreviousNext →