Font size
S
M
L
XL
Worksheetsmachine learning test
Total questions: 40
Worksheet time: 15mins
Name
Class
Date
1.
DecisionTreeRegressor ni to‘g‘ri import qilish:
a)
from sklearn.tree import DecisionTreeRegressor
b)
import DecisionTreeRegressor from sklearn
c)
from sklearn.model import DecisionTreeRegressor
d)
import sklearn.tree.DecisionTree
2.
Quyidagi koddan qaysi biri modelni o‘qitadi? (X_train, y_train mavjud)
a)
model = DecisionTreeRegressor(random_state=0); model.fit(X_train, y_train)
b)
model = DecisionTreeRegressor(); model.train(X_train, y_train)
c)
DecisionTreeRegressor.fit(X_train, y_train)
d)
fit(model, X_train, y_train)
3.
Bashorat natijasini olish uchun to‘g‘ri chaqiriq:
a)
preds = model.predict(X_valid)
b)
preds = model.transform(X_valid)
c)
preds = predict(model, X_valid)
d)
preds = model.forward(X_valid)
4.
Decision Tree chuqurligini cheklash parametri:
a)
max_depth
b)
n_estimators
c)
learning_rate
d)
alpha
5.
train_test_split’ni to‘g‘ri import qilish:
a)
from sklearn.model_selection import train_test_split
b)
from sklearn.split import train_test_split
c)
from sklearn.pipeline import train_test_split
d)
import train_test_split from sklearn
6.
Ma’lumotni 80/20 bo‘lib ajratish kodi:
a)
X_train, X_valid, y_train, y_valid = train_test_split(X, y, test_size=0.2, random_state=42)
b)
X_train, X_valid = train_test_split(X, y, 0.2)
c)
X_train, y_train, X_valid, y_valid = train_test_split(X, y, 0.8)
d)
X_train, X_valid, y_train, y_valid = split(X, y, 0.2)
7.
MAE (o‘rtacha mutlaq xatolik)ni hisoblash uchun import:
a)
from sklearn.metrics import mean_absolute_error
b)
from sklearn.metrics import mae_score
c)
from sklearn.losses import MAE
d)
import mae from sklearn.metrics
8.
MAE ni to‘g‘ri hisoblaydigan qator:
a)
mae = mean_absolute_error(y_valid, preds)
b)
mae = mean_absolute_error(preds)
c)
mae = MAE(y_valid, preds)
d)
mae = mean_absolute_error(X_valid, preds)
9.
5-fold kross-validatsiya uchun to‘g‘ri chaqiriq (MAE):
a)
scores = cross_val_score(model, X, y, cv=5, scoring='neg_mean_absolute_error')
b)
scores = cross_val_score(model, X, y, cv=5, scoring='mae')
c)
scores = cv_score(model, X, y, folds=5, metric='mae')
d)
scores = cross_val_score(model, X, y, cv=5) # klassifikatsiya default
10.
CV MAE o‘rtachasini to‘g‘ri olish:
a)
cv_mae = -scores.mean()
b)
cv_mae = scores.mean()
c)
cv_mae = abs(scores)
d)
cv_mae = mean_absolute_error(scores)
11.
RandomForestRegressor import:
a)
from sklearn.ensemble import RandomForestRegressor
b)
from sklearn.tree import RandomForestRegressor
c)
import RandomForestRegressor from sklearn
d)
from sklearn.forest import RandomForest
12.
RandomForestRegressor’da daraxtlar sonini belgilovchi parametr:
a)
n_estimators
b)
max_depth
c)
learning_rate
d)
gamma
13.
RandomForestRegressor bilan o‘qitish va bashorat:
a)
rf = RandomForestRegressor(random_state=0); rf.fit(X_train, y_train); preds = rf.predict(X_valid)
b)
rf = RandomForestRegressor(); rf.train(X_train, y_train); preds = rf.forward(X_valid)
c)
rf = RandomForestRegressor(); preds = rf.fit_predict(X_train, X_valid)
d)
rf = RandomForestClassifier(); rf.fit(...); rf.predict(...)
14.
OneHotEncoder import to‘g‘ri varianti:
a)
from sklearn.preprocessing import OneHotEncoder
b)
from sklearn.compose import OneHotEncoder
c)
from sklearn.pipeline import OneHotEncoder
d)
import OneHotEncoder from sklearn
15.
SimpleImputer import to‘g‘ri varianti:
a)
from sklearn.impute import SimpleImputer
b)
from sklearn.preprocessing import SimpleImputer
c)
from sklearn.fill import SimpleImputer
d)
import SimpleImputer from sklearn
16.
ColumnTransformer import to‘g‘ri varianti:
a)
from sklearn.compose import ColumnTransformer
b)
from sklearn.pipeline import ColumnTransformer
c)
from sklearn.column import Transformer
d)
from sklearn.compose import ColumnSelect
17.
OneHotEncoder’da noma’lum kategoriyalarni xatoliksiz o‘tish parametri:
a)
handle_unknown='ignore'
b)
unknown_ok=True
c)
missing_values='ignore'
d)
drop_unknown=True
18.
Raqamli va kategorik ustunlar uchun to‘g‘ri ColumnTransformer:
a)
preprocess = ColumnTransformer([
('num', SimpleImputer(strategy='median'), num_cols),
('cat', OneHotEncoder(handle_unknown='ignore'), cat_cols)
])
b)
preprocess = ColumnTransformer([
('num', OneHotEncoder(), num_cols),
('cat', SimpleImputer(), cat_cols)
])
c)
preprocess = ColumnTransformer(OneHotEncoder(), SimpleImputer())
d)
preprocess = ColumnTransformer([]) # bo‘sh
19.
Pipeline import to‘g‘ri varianti:
a)
from sklearn.pipeline import Pipeline
b)
from sklearn.compose import Pipeline
c)
from sklearn.pipeline import pipe
d)
import Pipeline from sklearn
20.
Preprocessing + modelni pipeline’da birlashtirish:
a)
pipe = Pipeline(steps=[('preprocess', preprocess), ('rf', RandomForestRegressor(random_state=0))])
b)
pipe = Pipeline([preprocess, RandomForestRegressor])
c)
pipe = make_pipeline(preprocess=preprocess, model=RandomForestRegressor())
d)
pipe = Pipeline(steps={'preprocess': preprocess, 'rf': RandomForestRegressor()})
21.
Pipeline’ni o‘qitishning to‘g‘ri yo‘li:
a)
pipe.fit(X_train, y_train)
b)
fit(pipe, X_train)
c)
pipe.train(X_train, y_train, X_valid)
d)
pipe.compile(X_train, y_train)
22.
Leakage’ni oldini olish uchun qaysi yondashuv to‘g‘ri?
a)
Imputer/Encoder’ni Pipeline ichida faqat train’da fit qilish
b)
Imputer’ni butun datasetga fit qilib, so‘ng train/testga qo‘llash
c)
Encoder’ni X va y birlashtirilgan holda fit qilish
d)
Validatsiyani train setga qo‘shib yuborish
23.
Pipeline bilan grid search param-nomi qanday beriladi? (Random Forest qadam nomi 'rf')
a)
param_grid = {'rf__n_estimators': [100, 200]}
b)
param_grid = {'n_estimators': [100, 200]}
c)
param_grid = {'preprocess.n_estimators': [100, 200]}
d)
param_grid = {'model__rf__n_estimators': [100, 200]}
24.
XGBRegressor import to‘g‘ri varianti:
a)
from xgboost import XGBRegressor
b)
from sklearn.xgboost import XGBRegressor
c)
import XGBRegressor from xgboost.sklearn
d)
from xgboost import XGBoostRegressor
25.
XGBRegressor bilan o‘qitish va bashorat:
a)
xgb = XGBRegressor(n_estimators=300, learning_rate=0.05, random_state=0); xgb.fit(X_train, y_train); preds = xgb.predict(X_valid)
b)
xgb = XGBRegressor(); preds = xgb.fit_predict(X_train, X_valid)
c)
xgb = XGBRegressor(); xgb.train(X_train, y_train); preds = xgb.forward(X_valid)
d)
xgb = XGBClassifier(); xgb.fit(X_train, y_train)
26.
XGBRegressor’da early stopping uchun to‘g‘ri chaqiriq:
a)
xgb.fit(X_train, y_train, eval_set=[(X_valid, y_valid)], early_stopping_rounds=50, verbose=False)
b)
xgb.fit(X_train, y_train, early_stop=True)
c)
xgb.fit(X_train, y_train, patience=50)
d)
xgb.train(..., early_stopping_rounds=50)
27.
cross_val_score bilan MAE uchun scoring nomi:
a)
'neg_mean_absolute_error'
b)
'mae'
c)
'mean_absolute_error'
d)
'abs_error'
28.
Pandas’da kategorik ustunlarni tanlashning to‘g‘ri yo‘li:
a)
cat_cols = X_train.select_dtypes(include=['object', 'category']).columns
b)
cat_cols = X_train.columns.category
c)
cat_cols = select_categorical(X_train)
d)
cat_cols = X_train.columns[object]
29.
OneHotEncoder o‘rniga pandas yordamida:
a)
X_enc = pd.get_dummies(X, columns=cat_cols)
b)
X_enc = pd.one_hot(X)
c)
X_enc = X.to_onehot()
d)
X_enc = pd.encode_dummies(X)
30.
make_column_selector’dan to‘g‘ri foydalanish:
a)
from sklearn.compose import make_column_selector as selector
b)
from sklearn.preprocessing import make_column_selector
c)
from sklearn.selector import make_column_selector
d)
from sklearn.compose import make_selector
31.
selector bilan ColumnTransformer misoli:
a)
ct = ColumnTransformer([
('num', SimpleImputer(strategy='median'), selector(dtype_include='number')),
('cat', OneHotEncoder(handle_unknown='ignore'), selector(dtype_include='object'))
])
b)
ct = ColumnTransformer([('all', SimpleImputer(), slice(None))])
c)
ct = ColumnTransformer(selector('all'))
d)
ct = ColumnTransformer([])
32.
Regression uchun to‘g‘ri metrikani tanlang:
a)
mean_absolute_error
b)
accuracy_score
c)
f1_score
d)
roc_auc_score
33.
Qayta tiklanuvchanlik (reproducibility) uchun odatiy amaliyot:
a)
random_state parametrini berish
b)
GPU ni o‘chirish
c)
always_shuffle=False
d)
CSV ni binar holatga o‘tkazish
34.
RandomForest’da murakkablikni cheklovchi parametr(lar):
a)
max_depth
b)
min_samples_leaf
c)
n_estimators
d)
max_features
35.
Qaysi holat DATA LEAKAGE ga misol?
a)
StandardScaler’ni butun datasetga fit qilib, so‘ng train/testga qo‘llash
b)
Skalerni Pipeline ichida fit qilish
c)
train_test_split’dan keyin fit qilish
d)
Only train’da fit, test’da faqat transform
36.
RandomForestRegressor’da barcha yadrolardan foydalanish parametri:
a)
n_jobs=-1
b)
workers='all'
c)
threads=0
d)
parallel=True
37.
train_test_split’da stratify=y qachon ishlatiladi?
a)
Klassifikatsiyada sinf ulushlarini saqlash uchun
b)
Har doim regressiyada
c)
XGBoost bilan shart
d)
Pipeline ishlaganda shart
38.
Kategorik ustunlarni 'missing' bilan to‘ldirish:
a)
SimpleImputer(strategy='constant', fill_value='missing')
b)
SimpleImputer(strategy='mean', fill_value='missing')
c)
SimpleImputer(fill='missing')
d)
Imputer(const='missing')
39.
ColumnTransformer’da tanlanmagan ustunlarni saqlab qolish uchun:
a)
remainder='passthrough'
b)
keep_others=True
c)
rest='keep'
d)
others='passthrough'
40.
Random Forest’da out-of-bag (OOB) baholashni yoqish:
a)
RandomForestRegressor(oob_score=True, bootstrap=True)
b)
RandomForestRegressor(oob=True)
c)
RandomForestRegressor(bootstrap=False, oob_score=True)
d)
RandomForestRegressor(score_oob=True)
Reset
