7.8 rain demo variable selection

Variable Selection

When the model was built, the algorithm chooses a variable for each
node of the resulting decision tree. An entropy, information theory or
gini based calculation is used to choose the variable. The variable
with the highest value according to this measure is chosen for the
particular node.

Below we will see the calculations that were made for the root node of
the tree (Node Number 1). A number of variables were considered and the
variable with the top score was chosen for this node. The improve= is
the value of the calculation.

Press Enter to continue: 

Node number 1: 123722 observations,    complexity param=0.342
  predicted class=no   expected loss=0.4  P(node) =1
    class counts: 97753 25969
   probabilities: 0.600 0.400 
  left son=2 (92796 obs) right son=3 (30926 obs)
  Primary splits:
      humidity_3pm < 64.5  to the left,  improve=11510, (0 missing)
      rainfall     < 0.35  to the left,  improve= 7486, (0 missing)
      rain_today   splits as  LR,        improve= 7133, (0 missing)
      cloud_3pm    < 6.5   to the left,  improve= 5030, (0 missing)
      humidity_9am < 73.5  to the left,  improve= 4535, (0 missing)
  Surrogate splits:
      cloud_3pm    < 7.5   to the left,  agree=0.778, adj=0.112, (0 split)
      humidity_9am < 87.5  to the left,  agree=0.775, adj=0.100, (0 split)
      sunshine     < 3.25  to the right, agree=0.772, adj=0.088, (0 split)
      temp_3pm     < 12.55 to the right, agree=0.771, adj=0.085, (0 split)
      rainfall     < 4.85  to the left,  agree=0.766, adj=0.063, (0 split)

Press Enter to continue: 

Your donation will support ongoing development and give you access to the PDF version of this book. Desktop Survival Guides include Data Science, GNU/Linux, and MLHub. Books available on Amazon include Data Mining with Rattle and Essentials of Data Science. Popular open source software includes rattle, wajig, and mlhub. Hosted by Togaware, a pioneer of free and open source software since 1984.
Copyright © 1995-2021 Graham.Williams@togaware.com Creative Commons Attribution-ShareAlike 4.0.