Steps

From WikiEducator
Jump to: navigation, search

Data mining is a process: steps


Data mining is not a single step process, neither a single goal process. Furthermore, data mining it's not a goal in itself (except for researchers), but a powerful tool for discovering knowledge hidden in the available data. In fact, "mining" is a good way to explain the whole process using the basic definition of the word "mining" [1], where the extracted minerals become the pursued knowledge, and the ground is the available data. Depending on the ground and the extracted mineral, several mining tools are available, thus it is necessary to develop a project for achieveing the desired goal.

  • Problem definition:
    • Discovering data similarities and patterns (clustering)
    • Classification
    • Prediction and estimation
    • Description and explanation
  • Data preprocessing:
  • Model construction
  • Evaluation and interpretation
  • Exploitation