Logistik regressiya va klassifikatsiya

import numpy as np
from sklearn.linear_model import LinearRegression

# 0 - imtihondan yiqildi, 1 - o'tdi
soatlar = np.array([1, 2, 3, 4, 5, 6, 7, 8, 20]).reshape(-1, 1)
otdi    = np.array([0, 0, 0, 0, 1, 1, 1, 1, 1])

model = LinearRegression().fit(soatlar, otdi)
print(model.predict([[10]]))
print(model.predict([[0]]))

def sigmoid(z):
    return 1 / (1 + np.exp(-z))

for z in [-5, -2, 0, 2, 5]:
    print(f"sigmoid({z:3d}) = {sigmoid(z):.4f}")

from sklearn.linear_model import LogisticRegression

model = LogisticRegression()
model.fit(soatlar, otdi)

print("Bashorat:", model.predict([[3], [6], [10]]))
print("Ehtimollik:")
print(model.predict_proba([[3], [6], [10]]).round(3))

import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.pipeline import make_pipeline
from sklearn.metrics import accuracy_score, classification_report

np.random.seed(42)
n = 500

talabalar = pd.DataFrame({
    "oqish_soati":   np.random.uniform(0, 20, n),
    "davomat":       np.random.uniform(40, 100, n),
    "oldingi_baho":  np.random.uniform(2, 5, n),
    "uyqu_soati":    np.random.uniform(4, 10, n),
})

# Yashirin qoida
ball = (
    talabalar["oqish_soati"] * 0.35
    + talabalar["davomat"] * 0.06
    + talabalar["oldingi_baho"] * 1.2
    + np.random.normal(0, 1.5, n)
)
talabalar["otdi"] = (ball > ball.median()).astype(int)

X = talabalar.drop(columns=["otdi"])
y = talabalar["otdi"]

X_o, X_s, y_o, y_s = train_test_split(
    X, y, test_size=0.2, random_state=42, stratify=y
)

model = make_pipeline(StandardScaler(), LogisticRegression())
model.fit(X_o, y_o)

bashorat = model.predict(X_s)
print(f"Aniqlik: {accuracy_score(y_s, bashorat):.3f}")

vaznlar = model.named_steps["logisticregression"].coef_[0]

for belgi, vazn in sorted(zip(X.columns, vaznlar), key=lambda p: -abs(p[1])):
    yonalish = "oshiradi" if vazn > 0 else "kamaytiradi"
    print(f"{belgi:15s} {vazn:6.2f}  o'tish ehtimolini {yonalish}")

ehtimollik = model.predict_proba(X_s)[:, 1]

# Standart chegara 0.5
bashorat_05 = (ehtimollik >= 0.5).astype(int)

# Qattiqroq chegara
bashorat_07 = (ehtimollik >= 0.7).astype(int)

print(f"0.5 chegara: {bashorat_05.sum()} ta 'o'tdi'")
print(f"0.7 chegara: {bashorat_07.sum()} ta 'o'tdi'")

from sklearn.datasets import load_iris

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

print("Sinflar:", iris.target_names)

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

model = make_pipeline(StandardScaler(), LogisticRegression(max_iter=1000))
model.fit(X_o, y_o)

print(f"Aniqlik: {model.score(X_s, y_s):.3f}")
print("Ehtimolliklar:")
print(model.predict_proba(X_s[:3]).round(3))

# C - regularizatsiya kuchining TESKARISI
for c in [0.01, 0.1, 1, 100]:
    model = make_pipeline(StandardScaler(), LogisticRegression(C=c, max_iter=1000))
    model.fit(X_o, y_o)
    print(f"C={c:6.2f}  o'qitish={model.score(X_o, y_o):.3f}  "
          f"sinov={model.score(X_s, y_s):.3f}")

import matplotlib.pyplot as plt

X_ikki = X[:, :2]     # faqat ikkita belgi
model = make_pipeline(StandardScaler(), LogisticRegression())
model.fit(X_ikki, y)

x_min, x_max = X_ikki[:, 0].min() - 0.5, X_ikki[:, 0].max() + 0.5
y_min, y_max = X_ikki[:, 1].min() - 0.5, X_ikki[:, 1].max() + 0.5

tor_x, tor_y = np.meshgrid(
    np.linspace(x_min, x_max, 300),
    np.linspace(y_min, y_max, 300),
)

Z = model.predict(np.c_[tor_x.ravel(), tor_y.ravel()]).reshape(tor_x.shape)

plt.figure(figsize=(8, 6))
plt.contourf(tor_x, tor_y, Z, alpha=0.25, cmap="viridis")
plt.scatter(X_ikki[:, 0], X_ikki[:, 1], c=y, cmap="viridis", edgecolors="k")
plt.xlabel(iris.feature_names[0])
plt.ylabel(iris.feature_names[1])
plt.title("Logistik regressiya qaror chegaralari")
plt.show()

from collections import Counter

# 95% sog'lom, 5% kasal
y_nomutanosib = np.array([0] * 950 + [1] * 50)
print(Counter(y_nomutanosib))

model_oddiy = LogisticRegression()
model_muvozanat = LogisticRegression(class_weight="balanced")