Qaror daraxti va Random Forest

from sklearn.tree import DecisionTreeClassifier, plot_tree
from sklearn.datasets import load_iris
from sklearn.model_selection import train_test_split
import matplotlib.pyplot as plt

iris = load_iris()
X, y = iris.data, iris.target

X_o, X_s, y_o, y_s = train_test_split(
    X, y, test_size=0.3, random_state=42, stratify=y
)

daraxt = DecisionTreeClassifier(max_depth=3, random_state=42)
daraxt.fit(X_o, y_o)

print(f"O'qitish: {daraxt.score(X_o, y_o):.3f}")
print(f"Sinov:    {daraxt.score(X_s, y_s):.3f}")

plt.figure(figsize=(16, 9))
plot_tree(
    daraxt,
    feature_names=iris.feature_names,
    class_names=iris.target_names,
    filled=True,
    rounded=True,
    fontsize=10,
)
plt.show()

from sklearn.tree import export_text

print(export_text(daraxt, feature_names=iris.feature_names))

def gini(sinflar):
    _, sonlar = np.unique(sinflar, return_counts=True)
    ehtimolliklar = sonlar / len(sinflar)
    return 1 - np.sum(ehtimolliklar ** 2)

print(gini([0, 0, 0, 0, 0]))              # 0.0
print(gini([0, 0, 0, 0, 1]))              # 0.32
print(gini([0, 0, 0, 1, 1]))              # 0.48
print(gini([0, 1, 0, 1, 0, 1]))           # 0.5

DecisionTreeClassifier(criterion="gini")       # standart
DecisionTreeClassifier(criterion="entropy")

daraxt_cheksiz = DecisionTreeClassifier(random_state=42)
daraxt_cheksiz.fit(X_o, y_o)

print(f"O'qitish: {daraxt_cheksiz.score(X_o, y_o):.3f}")
print(f"Sinov:    {daraxt_cheksiz.score(X_s, y_s):.3f}")
print(f"Chuqurlik: {daraxt_cheksiz.get_depth()}")
print(f"Barglar:   {daraxt_cheksiz.get_n_leaves()}")

daraxt = DecisionTreeClassifier(
    max_depth=4,              # maksimal chuqurlik
    min_samples_split=20,     # bo'linish uchun minimal namuna
    min_samples_leaf=10,      # bargdagi minimal namuna
    max_leaf_nodes=15,        # maksimal barglar soni
    min_impurity_decrease=0.01,
    random_state=42,
)

print("Chuqurlik  O'qitish  Sinov")
for chuqurlik in [1, 2, 3, 5, 10, None]:
    d = DecisionTreeClassifier(max_depth=chuqurlik, random_state=42).fit(X_o, y_o)
    nom = str(chuqurlik) if chuqurlik else "cheksiz"
    print(f"{nom:>9}  {d.score(X_o, y_o):8.3f}  {d.score(X_s, y_s):.3f}")

import pandas as pd

daraxt = DecisionTreeClassifier(max_depth=3, random_state=42).fit(X_o, y_o)

muhimlik = pd.Series(
    daraxt.feature_importances_,
    index=iris.feature_names,
).sort_values(ascending=False)

print(muhimlik.round(3))

from sklearn.ensemble import RandomForestClassifier

orman = RandomForestClassifier(
    n_estimators=200,
    max_depth=None,
    max_features="sqrt",
    n_jobs=-1,
    random_state=42,
)
orman.fit(X_o, y_o)

print(f"Bitta daraxt: {daraxt_cheksiz.score(X_s, y_s):.3f}")
print(f"Random Forest: {orman.score(X_s, y_s):.3f}")

orman = RandomForestClassifier(
    n_estimators=200, oob_score=True, random_state=42, n_jobs=-1,
)
orman.fit(X_o, y_o)

print(f"OOB aniqlik:   {orman.oob_score_:.3f}")
print(f"Sinov aniqlik: {orman.score(X_s, y_s):.3f}")

from sklearn.ensemble import GradientBoostingClassifier, HistGradientBoostingClassifier

gb = GradientBoostingClassifier(
    n_estimators=100,
    learning_rate=0.1,
    max_depth=3,
    random_state=42,
)
gb.fit(X_o, y_o)
print(f"Gradient Boosting: {gb.score(X_s, y_s):.3f}")

# Katta ma'lumot uchun ancha tezroq
hgb = HistGradientBoostingClassifier(random_state=42).fit(X_o, y_o)
print(f"HistGradientBoosting: {hgb.score(X_s, y_s):.3f}")

from sklearn.ensemble import RandomForestRegressor

model = RandomForestRegressor(n_estimators=200, n_jobs=-1, random_state=42)
model.fit(X_o, y_o)