Decision Tree from Scratch

• This version of Decision Tree is completly build using algorithm ID3 (Iterative Dichotomiser 3).

• In code Mostly work is done using Pandas and math

  • Pandas for Data manipulation.
  • math for mathematical operations.

• It's also Construct the Tree using Node class.

• One simple visilization of tree is also there. (if verbose = 1)

     Outlook (Gain: 0.2467)
    |── Rainy
    Windy (Gain: 0.9403)
        |── Strong
        |── Weak
    |── Overcast
    |── Sunny
    Humidity (Gain: 0.9403)
        |── High
        |── Normal 
  

• Example data:

  • Train Data : Outlook, Temperature, Humidity, Windy
  • Label : PlayTennis

  Outlook	Temperature	Humidity	Windy	PlayTennis
  Sunny	Hot	High	Weak	No
  Sunny	Hot	High	Strong	No
  Overcast	Hot	High	Weak	Yes
  Rainy	Mild	High	Weak	Yes
  Rainy	Cool	Normal	Weak	Yes
  Rainy	Cool	Normal	Strong	No
  Overcast	Cool	Normal	Strong	Yes
  Sunny	Mild	High	Weak	No
  Sunny	Cool	Normal	Weak	Yes
  Rainy	Mild	Normal	Weak	Yes
  Sunny	Mild	Normal	Strong	Yes
  Overcast	Mild	High	Strong	Yes
  Overcast	Hot	High	Weak	Yes
  Rainy	Mild	High	Strong	No