OpenEDG PCEI — Certified Entry-Level AI Engineer with Python

Which statement BEST describes the relationship between AI, ML, and deep learning?

• AThey are synonyms
• BDeep learning is a subset of machine learning, which is a subset of artificial intelligen…
• CAI is a subset of machine learning
• DDeep learning and AI are both subsets of machine learning

Which type of machine learning learns from labeled examples?

• ASupervised learning
• BUnsupervised learning
• CReinforcement learning
• DSemi-supervised learning

Which task is an example of REGRESSION (not classification)?

• APredicting whether an email is spam
• BPredicting tomorrow's temperature in Celsius
• CClassifying images as cat or dog
• DIdentifying handwritten digits 0-9

Which task is an example of CLUSTERING?

• APredicting customer churn given labeled data
• BGrouping customers into segments based on purchasing behavior without labels
• CPredicting house prices
• DTranslating English to French

What is reinforcement learning's main feedback signal?

• ALabels
• BReward (and penalty)
• CPre-existing clusters
• DLoss function gradient only

What does 'training a model' mean in machine learning?

• AAdjusting model parameters to minimize a loss function on training data
• BRunning the model on new data
• CCleaning the input data
• DTesting the deployed model

Why do we split data into train, validation, and test sets?

• ATo make training faster
• BTo estimate generalization, tune hyperparameters without leakage, and report unbiased fin…
• CTo use less memory
• DBecause the law requires it

What is overfitting?

• AWhen the model performs poorly on both training and test data
• BWhen the model memorizes training data and fails to generalize to new data
• CWhen training takes too long
• DWhen the model is too simple

What is underfitting?

• AThe model is too complex for the data
• BThe model is too simple to capture the underlying pattern, performing poorly on both trai…
• CThe model overfits to noise
• DTraining data is too large

Which is NOT a typical step in a machine learning pipeline?

• AData collection
• BData preprocessing and feature engineering
• CCompiling Python source to assembly
• DModel training and evaluation

Which import statement is the conventional alias for NumPy?

• Aimport numpy
• Bimport numpy as np
• Cfrom numpy import *
• Dimport numpy as numpy

How do you create a NumPy array of zeros with shape (3, 4)?

• Anp.zeros(3, 4)
• Bnp.zeros((3, 4))
• Cnp.empty(3, 4)
• Dnp.array.zeros(3, 4)

What is broadcasting in NumPy?

• ABroadcasting is sending arrays over a network
• BAutomatic expansion of smaller arrays to match larger arrays' shapes for elementwise oper…
• CPrinting arrays in a formatted way
• DCompressing arrays

What does this code print? ```python import numpy as np a = np.array([1, 2, 3, 4]) print(a.mean()) ```

• A1
• B2.5
• C10
• D4

How do you create a pandas DataFrame from a dictionary?

• Apd.frame({'a': [1,2], 'b': [3,4]})
• Bpd.DataFrame({'a': [1,2], 'b': [3,4]})
• Cpd.dict_to_frame({'a': [1,2], 'b': [3,4]})
• Dpd.from_dict({'a': [1,2], 'b': [3,4]}, orient='rows')

Which DataFrame method handles missing values by removing rows containing them?

• Adf.fillna()
• Bdf.dropna()
• Cdf.remove()
• Ddf.clean()

What does df.iloc[0:5, 1:3] return?

• AThe first 5 columns and rows 1-3
• BRows at positions 0-4 and columns at positions 1-2 (exclusive end)
• CA scalar value at position [0:5, 1:3]
• DA copy of the entire DataFrame

Which sklearn preprocessing class scales features to zero mean and unit variance?

• AMinMaxScaler
• BStandardScaler
• CRobustScaler
• DNormalizer

What does this code do? ```python from sklearn.preprocessing import OneHotEncoder enc = OneHotEncoder() X_encoded = enc.fit_transform(X_categorical) ```

• AConverts numeric features to categories
• BConverts categorical variables to one-hot encoded binary columns
• CStandardizes the features
• DRemoves outliers

What is the typical sklearn function for splitting data into train and test sets?

• Asklearn.split(X, y, ratio=0.2)
• Bsklearn.model_selection.train_test_split(X, y, test_size=0.2)
• Csklearn.preprocessing.split(X, y)
• Dsklearn.utils.holdout(X, y)

Why is it bad to fit a StandardScaler on the FULL dataset before splitting into train/test?

• AIt is computationally expensive
• BIt causes data leakage — test set statistics inform the scaling, biasing evaluation
• CIt doesn't work in sklearn
• DIt only works for classification

Which sklearn class chains preprocessing and a model into a single estimator?

• Asklearn.chain.Chain
• Bsklearn.pipeline.Pipeline
• Csklearn.combine.Combine
• Dsklearn.preprocessing.Sequence

Which sklearn function performs k-fold cross-validation and returns scores?

• Asklearn.evaluation.cross_validate
• Bsklearn.model_selection.cross_val_score
• Csklearn.metrics.k_fold
• Dsklearn.cv.score

Which Pandas method merges two DataFrames on a common column?

• Adf1.merge(df2, on='id')
• Bpd.combine(df1, df2)
• Cdf1.append(df2)
• Ddf1 + df2

What is the typical NumPy dtype for floating-point ML computations?

• Aint32
• Bfloat64
• Cstring
• Dobject

How do you reshape a NumPy array of 12 elements into a 3x4 matrix?

• Aa.shape = 3, 4
• Ba.reshape(3, 4)
• Ca.resize((3, 4))
• DAll of the above can work, but a.reshape(3, 4) is the conventional choice

Which sklearn class implements a logistic regression classifier?

• Asklearn.linear_model.LogisticRegression
• Bsklearn.classifiers.Logistic
• Csklearn.regression.Logistic
• Dsklearn.tree.LogisticRegression

What does this code do? ```python from sklearn.ensemble import RandomForestClassifier clf = RandomForestClassifier(n_estimators=100) clf.fit(X_train, y_train) preds = clf.predict(X_test) ```

• ATrains a single decision tree
• BTrains an ensemble of 100 decision trees and predicts on the test set
• CPerforms k-means clustering
• DComputes feature correlations

Which sklearn class implements k-Nearest Neighbors classification?

• AKNeighborsClassifier
• BNearestNeighborClf
• CKNNClassifier
• DKMeansClassifier

Which classification metric is the harmonic mean of precision and recall?

• AAccuracy
• BF1 score
• CROC-AUC
• DMean squared error

Which sklearn function computes the confusion matrix?

• Asklearn.metrics.confusion_matrix(y_true, y_pred)
• Bsklearn.metrics.matrix(y_true, y_pred)
• Csklearn.confusion(y_true, y_pred)
• Dsklearn.evaluate.confusion(y_true, y_pred)

What is the difference between precision and recall?

• AThey are the same
• BPrecision = TP/(TP+FP); Recall = TP/(TP+FN)
• CPrecision uses true labels; recall uses predicted labels
• DPrecision is for regression; recall is for classification

Which sklearn class performs k-means clustering?

• AKMeans
• BKMeansClustering
• CClusterKMeans
• DCluster

Which density-based clustering algorithm does NOT require specifying the number of clusters?

• AKMeans
• BDBSCAN
• CGaussianMixture
• DAgglomerativeClustering

Which metric is appropriate for evaluating a REGRESSION model?

• Aaccuracy_score
• Bmean_squared_error
• Cf1_score
• Dconfusion_matrix

What does sklearn's GridSearchCV do?

• APerforms k-fold cross-validation only
• BSearches over a grid of hyperparameters using cross-validation to find the best combinati…
• CPlots the model's decision boundary
• DGenerates a confusion matrix

What is regularization in machine learning?

• ARemoving features with zero variance
• BAdding a penalty to the loss function to prevent overfitting (e.g., L1 or L2 norm of weig…
• CNormalizing the input data
• DSplitting data into folds

Which sklearn dataset loader returns the iris dataset?

• Aload_iris()
• Biris_data()
• Cfetch_iris()
• Ddatasets.iris

Which sklearn class implements a linear regression model?

• Asklearn.linear_model.LinearRegression
• Bsklearn.regression.Linear
• Csklearn.models.Linear
• Dsklearn.tree.LinearRegressor

What does .fit_transform() do in a sklearn transformer?

• AFits and predicts in one call
• BFits the transformer on data, then transforms that same data — equivalent to .fit(X).tran…
• CTrains a model and evaluates it
• DReverses a previous transformation

Which sklearn metric returns R-squared (coefficient of determination)?

• Amean_squared_error
• Br2_score
• Caccuracy_score
• Dlog_loss

What does class_weight='balanced' do in sklearn classifiers?

• ARemoves minority class samples
• BAutomatically weights classes inversely proportional to their frequencies, helping with i…
• CDoubles the dataset
• DComputes balanced accuracy

Which is TRUE about Decision Trees in sklearn?

• AThey cannot handle categorical features
• BThey are prone to overfitting; use max_depth or min_samples_leaf to regularize
• CThey require feature scaling
• DThey cannot do multi-class classification

Which activation function outputs values between 0 and 1, often used for binary classification output layers?

• AReLU
• BSigmoid
• CTanh
• DLinear

Which activation function is the most common default for hidden layers in modern deep networks?

• ASigmoid
• BTanh
• CReLU
• DStep function