Rule Based Machine Learning Overview

Definition of Machine Learning

• More experience -> more prior knowledge -> imporve experience
• The effort of producing a better approximate function

Perfect World for Rule Based Learning

Example

• Perfect world
  - no observation error
  • no stochastic elements in the system we observe
  • Full information in the observations to regenerate the system

Function Approximation

• instance X:
  • Features: sunny, warm, normal, strong, warm, same
  • Label: yes
• Hypotheses: H
  • $h_i$: <sunny, warm, ?, ?, ?, same> => yes
• Target function: C
  • unknown target function between features and label
• Graphical Representation