35. Decision Trees and Random Forests in Python
35.2 One Tree vs Random Forest
Why. One tree is easy to read but can be unstable. A random forest grows many trees on random row/feature slices and averages their votes.
What. Forest ≈ “committee of trees.” Usually stronger and less sensitive to one weird row; harder to draw on a whiteboard than a single small tree.
How.
When to prefer which
- Need explainable classroom rules → start with one shallow tree.
- Need better accuracy on tabular India ops data → try RandomForest.
- Still split train/test either way.
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import RandomForestClassifier
# Same X, y idea as Chapter 34 — forest is many trees voting
print(DecisionTreeClassifier)
print(RandomForestClassifier)
Ravindra Bagale's Tip
Forest = jungal of trees. Ek tree explain; forest performance. Interview madhe fark sanga.