Data Mining · CSC213

The KDD Process: Knowledge Discovery in Databases Steps

Data Mining Study guides

KDD (knowledge discovery in databases) is the overall process of turning raw data into knowledge; data mining is the pattern-finding step inside it.

The usual KDD steps run in order: data cleaning (remove noise and inconsistent data), data integration (combine multiple data sources), data selection (retrieve the data relevant to the analysis), data transformation (transform or consolidate the data into forms appropriate for mining, for example by summary or aggregation), data mining (apply intelligent methods to extract patterns), pattern evaluation (identify the truly interesting patterns), and knowledge presentation (show the knowledge to the user with visualization).

The first steps are all preparation. Cleaning and integration usually take most of the effort in a real project, because mining algorithms are only as good as the data they get. Data mining itself is one step of the process, although people often use the term for the whole thing.

Data warehouses often sit behind the integration step: several sources are combined into one consistent store before selection.

Key points

  • Order: cleaning, integration, selection, transformation, mining, evaluation, presentation.
  • Integration = combining multiple data sources.
  • Transformation = consolidating data into forms appropriate for mining.
  • Data mining is one step of KDD.
  • Pattern evaluation keeps only the interesting patterns.

Worked example

A retailer has sales in one database and customer data in another. It cleans typos and duplicates, integrates both on customer ID, selects the last year, aggregates purchases per customer, mines for segments, evaluates which segments are useful, and shows the result as charts.

Common questions

What does KDD stand for?

Knowledge discovery in databases.

Is data mining the same as KDD?

Data mining is the core step of the KDD process, not the whole of it.

Practice questions on this topic

Try these, then open the app to check your answers with explanations.

  1. Multiple numbers of data sources get combined in which step of the Knowledge Discovery?
    • A. Data Cleaning
    • B. Data Integration
    • C. Data Selection
    • D. Data Transformation
  2. The knowledge discovery process in which several data are combined ______.
    • A. Data cleaning
    • B. Data integration
    • C. Date selection
    • D. None of the above
  3. Where data are transformed or consolidated into forms appropriate for mining by performing summary or aggregation operations
    • A. Data transformation
    • B. Data mining
    • C. Data knowledge
    • D. Data integration
  4. The initial steps concerned in the process of knowledge discovery is:
    • A. Data Transformation
    • B. Knowledge presentation
    • C. Data Integration
    • D. Data Cleaning
  5. The correct order in which all sub-processes of data mining execute is:
    • A. Infrastructure, Preprocessing, data mining, evaluation, Knowledge presentation.
    • B. Infrastructure, Preprocessing, Knowledge presentation, data mining, evaluation
    • C. data mining, Preprocessing, evaluation, Knowledge presentation
  6. Which of the following is NOT a step in the typical Knowledge Discovery (KDD) process?
    • A. Post-Processing (Pattern evaluation, visualization)
    • B. Data Pre-Processing
    • C. Data Mining
    • D. Initial Data Generation