Baza (baseline) va birinchi model

import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.dummy import DummyClassifier
from sklearn.metrics import accuracy_score, f1_score

df = pd.read_csv(DATA / "uy_ijara.csv")
son = ["xonalar", "maydon_m2", "qavat", "qavatlar_soni", "metro_km", "mebel", "qurilgan_yil"]
X, y = df[son + ["tuman"]], df["ijaraga_berildi_30k"]
X_tr, X_te, y_tr, y_te = train_test_split(X, y, test_size=0.3, random_state=42, stratify=y)

print(f"Target nisbati: {y.mean():.1%} 'ha'")

strategiyalar = {
    "most_frequent": "Har doim eng ko'p uchraydigan sinfni aytadi",
    "stratified": "Sinf nisbatiga qarab tasodifiy aytadi",
    "uniform": "Tanga tashlaydi (teng ehtimol)",
    "prior": "most_frequent bilan bir xil bashorat qiladi",
}
qatorlar = []
for s, izoh in strategiyalar.items():
    d = DummyClassifier(strategy=s, random_state=42).fit(X_tr[son], y_tr)
    b = d.predict(X_te[son])
    qatorlar.append({"Strategiya": s, "Accuracy": round(accuracy_score(y_te, b), 4),
                     "F1": round(f1_score(y_te, b, zero_division=0), 4), "Nima qiladi": izoh})
show(pd.DataFrame(qatorlar))

baza = accuracy_score(y_te, DummyClassifier(strategy="most_frequent")
                      .fit(X_tr[son], y_tr).predict(X_te[son]))
print(f"BAZA = {baza:.4f}")
print()
print("Shu raqamni daftaringizga yozing. Har bir keyingi natija shunga taqqoslanadi.")

from sklearn.pipeline import Pipeline
from sklearn.compose import ColumnTransformer
from sklearn.preprocessing import StandardScaler, OneHotEncoder
from sklearn.impute import SimpleImputer
from sklearn.linear_model import LogisticRegression

quvur = Pipeline([
    ("tayyorla", ColumnTransformer([
        ("son", Pipeline([("toldir", SimpleImputer(strategy="median")),
                          ("shkala", StandardScaler())]), son),
        ("matn", OneHotEncoder(handle_unknown="ignore"), ["tuman"]),
    ])),
    ("model", LogisticRegression(max_iter=2000)),
]).fit(X_tr, y_tr)

model_acc = accuracy_score(y_te, quvur.predict(X_te))
print(f"BAZA   : {baza:.4f}")
print(f"MODEL  : {model_acc:.4f}")
print(f"Farq   : {model_acc - baza:+.4f}")
print()
print("Hukm:", "O'TDI — model haqiqatan o'rgandi" if model_acc > baza + 0.02 else "O'TMADI")

td = pd.read_csv(DATA / "talaba_bitirdi.csv", keep_default_na=False, na_values=[""])
th = ["kirish_bali", "davomat_foiz", "uy_vazifa_bali", "ishlaydi", "masofa_km"]
Xt, yt = td[th], td["bitirdi"]
Xt_tr, Xt_te, yt_tr, yt_te = train_test_split(Xt, yt, test_size=0.3, random_state=42, stratify=yt)

baza_t = accuracy_score(yt_te, DummyClassifier(strategy="most_frequent")
                        .fit(Xt_tr, yt_tr).predict(Xt_te))
model_t = LogisticRegression(max_iter=2000).fit(Xt_tr, yt_tr)
acc_t = accuracy_score(yt_te, model_t.predict(Xt_te))

print(f"Target nisbati : {yt.mean():.1%} bitirgan")
print(f"BAZA           : {baza_t:.4f}")
print(f"MODEL          : {acc_t:.4f}")
print(f"Farq           : {acc_t - baza_t:+.4f}")

print("Xuddi shu model, boshqa metrikalar bilan:")
b = model_t.predict(Xt_te)
print(f"  Accuracy : {accuracy_score(yt_te, b):.4f}   (baza {baza_t:.4f} — foydasi kam)")
print(f"  F1       : {f1_score(yt_te, b):.4f}")
from sklearn.metrics import recall_score, precision_score
print(f"  Precision: {precision_score(yt_te, b, zero_division=0):.4f}")
print(f"  Recall   : {recall_score(yt_te, b, zero_division=0):.4f}")

from sklearn.dummy import DummyRegressor
from sklearn.linear_model import Ridge
from sklearn.metrics import root_mean_squared_error, r2_score

yn = df["narx_usd"]
Xn_tr, Xn_te, yn_tr, yn_te = train_test_split(X, yn, test_size=0.3, random_state=42)

d_reg = DummyRegressor(strategy="mean").fit(Xn_tr[son], yn_tr)
rmse_baza = root_mean_squared_error(yn_te, d_reg.predict(Xn_te[son]))

quvur_reg = Pipeline([
    ("tayyorla", ColumnTransformer([
        ("son", Pipeline([("toldir", SimpleImputer(strategy="median")),
                          ("shkala", StandardScaler())]), son),
        ("matn", OneHotEncoder(handle_unknown="ignore"), ["tuman"]),
    ])),
    ("model", Ridge()),
]).fit(Xn_tr, yn_tr)
bashorat_reg = quvur_reg.predict(Xn_te)

print(f"BAZA  (har doim o'rtacha): RMSE = {rmse_baza:.1f}  R2 = {r2_score(yn_te, d_reg.predict(Xn_te[son])):.4f}")
print(f"MODEL (Ridge)            : RMSE = {root_mean_squared_error(yn_te, bashorat_reg):.1f}  R2 = {r2_score(yn_te, bashorat_reg):.4f}")

show(pd.DataFrame([
    {"Baza turi": "Dummy (most_frequent)", "Nimani ko'rsatadi": "Model umuman o'rganganmi",
     "Kim so'raydi": "Siz"},
    {"Baza turi": "Oddiy qoida (if narx < X)", "Nimani ko'rsatadi": "ML murakkabligi kerakmi",
     "Kim so'raydi": "Senior muhandis"},
    {"Baza turi": "Hozirgi jarayon (odam/eski tizim)", "Nimani ko'rsatadi": "Loyiha foyda keltiradimi",
     "Kim so'raydi": "Menejer — eng muhim savol"},
]))

import numpy as np
chegara = X_tr["maydon_m2"].median()
qoida = (X_te["maydon_m2"] < chegara).astype(int)   # kichik kvartira tezroq ketadi degan taxmin
print(f"Qoida: maydon < {chegara:.0f} m2 bo'lsa 'tez ketadi'")
print(f"  Accuracy: {accuracy_score(y_te, qoida):.4f}")
print(f"  F1      : {f1_score(y_te, qoida, zero_division=0):.4f}")
print()
print(f"Taqqoslash -> dummy {baza:.4f} | qoida {accuracy_score(y_te, qoida):.4f} | model {model_acc:.4f}")

show(pd.DataFrame([
    {"Tushuncha": "Baza nima uchun", "Bitta jumla": "Busiz accuracy raqami ma'nosiz"},
    {"Tushuncha": "DummyClassifier", "Bitta jumla": "most_frequent — eng qattiq chegara"},
    {"Tushuncha": "DummyRegressor", "Bitta jumla": "Har doim o'rtacha; R2=0 shu nuqta"},
    {"Tushuncha": "Yuqori baza", "Bitta jumla": "0.85 dan oshsa accuracy ni tashlab F1 ga o't"},
    {"Tushuncha": "Oddiy qoida bazasi", "Bitta jumla": "ML umuman kerakmi degan savolga javob"},
    {"Tushuncha": "Biznes bazasi", "Bitta jumla": "Hozir buni kim qilyapti va qanchalik yaxshi"},
]))