MLA-C01 Practice Test 13 — 15 Questions

Question 1

An ML team is using Amazon SageMaker Feature Store to serve features for both real-time inference and batch training. They need to ensure that training data uses feature values as they were at the time of each event. Which type of query should they use?

Choose an answer to begin — your selection is scored in the full session.

Accepted Answer

Point-in-time query from the offline store. Point-in-time queries retrieve the correct feature values as of a specific timestamp, ensuring consistency between training and inference.

Answer

Full table scan of the offline store

Answer

Join query across both stores

Answer

Latest record query from the online store

Question 2

A data scientist needs to prepare a dataset for a binary classification model. The dataset has 10,000 samples with 100 features, and the target class is highly imbalanced (5% positive). Which combination of techniques should the data scientist use to address class imbalance and prepare the data for training?

Accepted Answer

Apply SMOTE to oversample the minority class, then use stratified splitting for train/test sets.. SMOTE generates synthetic samples for the minority class, and stratified splitting ensures the same class proportion in train/test sets. Lasso regression can help with feature selection but is not directly for imbalance. Undersampling would discard too many majority samples. Random splitting without stratification could lead to uneven class distribution.

Answer

Use random undersampling of the majority class, then apply MinMaxScaler to all features.

Answer

Use RandomOverSampler to duplicate minority samples, then perform a random train-test split.

Answer

Apply Lasso regression for feature selection, then use k-fold cross-validation without stratification.

Question 3

A machine learning team is preparing a dataset for training a deep learning model. They notice that some features have very different scales: one feature ranges from 0 to 1, another from 0 to 100,000, and a third is a binary indicator (0/1). The model uses gradient descent. Which scaling method should be applied to ALL features to ensure stable and efficient training?

Accepted Answer

StandardScaler to standardize all features to zero mean and unit variance. StandardScaler (z-score standardization) centers features to mean 0 and unit variance. It is robust to outliers (compared to MinMaxScaler which is sensitive to extreme values) and works well with gradient descent when features have varying scales.

Answer

MinMaxScaler to scale all features to [0,1]

Answer

RobustScaler to scale based on percentiles

Answer

No scaling is needed because gradient descent can handle different scales

Question 4

A data scientist is preparing a dataset for a multi-class classification problem. The dataset contains a categorical feature with 50,000 unique values (high cardinality). The scientist wants to reduce dimensionality while preserving predictive information. Which TWO approaches are appropriate? (Choose 2)

Accepted Answer

Target encoding. Target encoding replaces high-cardinality categories with the target mean, compressing the feature. Count encoding replaces categories with frequency counts. Both reduce dimensionality to one column. One-hot encoding would create 50,000 columns. Label encoding imposes order. Feature hashing can also work but is less common; count encoding is a valid option.

Answer

One-hot encoding

Answer

Ordinal encoding based on alphabetical order

Answer

Label encoding

Question 5

A machine learning engineer is preparing a dataset for a binary classification model. The dataset has 10,000 samples with a 1:100 class imbalance. The engineer needs to balance the classes before training. Which technique would create a balanced dataset without discarding majority class samples and without generating synthetic data?

Accepted Answer

Random oversampling of the minority class. Random oversampling duplicates minority class samples until classes are balanced. It does not discard majority samples (unlike undersampling) and does not generate synthetic data (unlike SMOTE).

Answer

Cost-sensitive learning with class weights

Answer

Synthetic Minority Over-sampling Technique (SMOTE)

Answer

Random undersampling of the majority class

Question 6

A financial services company needs to build a fraud detection model using historical transaction data. The dataset has a timestamp column, and the model must be evaluated on its ability to detect fraud in future unseen transactions. The data is imbalanced (fraud is rare). Which TWO data splitting strategies should the engineer use for model validation? (Select TWO.)

Accepted Answer

Stratified k-fold cross-validation. Time-series split (walk-forward validation) respects temporal order, and stratified k-fold cross-validation maintains class proportions across folds. Simple random split ignores time; k-fold without stratification may produce folds without fraud.

Answer

k-fold cross-validation without stratification

Answer

Leave-one-out cross-validation

Answer

Simple random split

Question 7

A data scientist is working with a dataset that contains missing values in several numerical columns. The missing data is not completely at random (MNAR). The scientist wants to minimize bias in the imputed values. Which imputation strategy is most appropriate?

Accepted Answer

Use a model-based imputation like iterative imputer or KNN imputer. Model-based imputation methods, such as using a regressor to predict missing values based on other features, can capture complex relationships and reduce bias compared to simple mean/median imputation, especially when data is not MCAR.

Answer

Delete all rows with missing values

Answer

Replace missing values with the median of each column

Answer

Replace missing values with the mean of each column

Question 8

A data scientist is preparing a dataset with a categorical feature that has 20 levels. The target variable is continuous. Which THREE encoding methods are appropriate for this scenario? (Select THREE.)

Accepted Answer

One-hot encoding. One-hot encoding, ordinal encoding (if order exists), and target encoding are all applicable for categorical features with a continuous target. Label encoding is similar to ordinal but usually implies arbitrary order, but still acceptable; however, the question expects three of the listed. The three most directly appropriate are one-hot, ordinal, and target.

Answer

Binary encoding

Answer

Label encoding

Question 9

A data scientist is using Amazon SageMaker Data Wrangler to create a feature engineering pipeline for a dataset with both numeric and categorical features. The scientist wants to apply transformations that are appropriate for a linear model. Which THREE transformations should the scientist apply? (Choose THREE.)

Accepted Answer

Remove features with high pairwise correlation. Linear models assume features are numeric, scaled, and not highly correlated. StandardScaler ensures all numeric features have comparable scales. One-hot encoding converts categorical features into binary columns without imposing ordinality. Removing highly correlated features reduces multicollinearity which can destabilize coefficient estimates.

Answer

MinMaxScaler on numeric features

Answer

Label encoding on categorical features

Question 10

A data engineer is using Amazon SageMaker Ground Truth to create a labeled dataset for an object detection task. The dataset contains millions of images, and the labeling budget is limited. Which approach can reduce labeling costs while maintaining high model accuracy?

Accepted Answer

Enable active learning in Ground Truth to automatically select a subset of images for human labeling. Active learning in Ground Truth selects the most informative samples (e.g., uncertain predictions) for human labeling, reducing the number of labels needed while maximizing model improvement. This is a built-in feature.

Answer

Label 100% of the images using a pre-built worker template to ensure accuracy

Answer

Use Amazon SageMaker Data Wrangler to annotate images

Answer

Use automated labeling with a pre-trained model for all images and skip human review

Question 11

A data scientist is building a regression model to predict house prices. The dataset contains a feature 'neighborhood' with 500 distinct values, and most neighborhoods have fewer than 10 samples. Which approach is MOST appropriate for handling this high-cardinality categorical feature?

Accepted Answer

Use target encoding with proper cross-validation to avoid data leakage. Target encoding replaces each category with the mean target value, which is effective for high-cardinality features while maintaining predictive power. One-hot encoding would create too many sparse columns, and label encoding would impose an arbitrary ordinal relationship.

Answer

Drop the neighborhood feature entirely

Answer

Apply frequency encoding, replacing each neighborhood with its count in the training set

Answer

One-hot encode the feature and use L1 regularization

Question 12

A financial services company is building a fraud detection model using historical transaction data stored in Amazon S3. The data includes features such as transaction amount, merchant category, time of day, and user location. The data scientist observes that the 'merchant_category' column is a text attribute with over 200 unique values. Additionally, the 'transaction_amount' column has a long-tail distribution with extreme outliers. The dataset is 200 GB in size, and the company wants to use Amazon SageMaker for model training. The data scientist needs to engineer features that capture the high-cardinality category and reduce the impact of outliers. What is the MOST efficient and effective approach to prepare this data?

Accepted Answer

Use Amazon EMR with Spark to apply ordinal encoding to merchant_category based on frequency, and log-transform the transaction_amount to reduce skewness.. Ordinal encoding based on frequency handles high-cardinality categorical features efficiently without exploding dimensionality, and log-transform is a standard technique to reduce skewness in long-tail distributions. Using Amazon EMR with Spark provides distributed processing for the 200 GB dataset, making it scalable and cost-effective compared to single-node alternatives.

Answer

Use AWS Glue ETL to apply one-hot encoding to merchant_category and min-max scaling to transaction_amount.

Answer

Use Amazon Athena to bin transaction_amount into 10 equal-width bins and replace merchant_category with its count encoding.

Answer

Use AWS Glue DataBrew to apply a one-hot encoding on merchant_category and a standard scaler on transaction_amount after removing outliers.

Question 13

A data scientist wants to compare the performance of two model versions (V1 and V2) in production by splitting traffic between them. They want to gradually increase the percentage of traffic to the new version while monitoring metrics. Which SageMaker feature enables this?

Accepted Answer

SageMaker production variants with traffic splitting. Production variants with traffic splitting allow routing a percentage of inference requests to different model versions. By updating the initial variant weights, the data scientist can gradually shift traffic from V1 to V2. Shadow testing mirrors traffic but does not affect real responses. Blue/green is a deployment pattern but not a SageMaker feature for gradual traffic splitting. Canary deployments are a pattern but SageMaker implements it via production variants.

Answer

SageMaker shadow testing

Answer

SageMaker blue/green deployment

Answer

SageMaker canary deployment

Question 14

A team wants to deploy a new model using a canary deployment strategy on SageMaker. Which TWO configurations are necessary? (Choose two.)

Accepted Answer

Set the initial traffic distribution (e.g., 90% old, 10% new). A canary deployment requires two production variants (old and new) with traffic distribution. Gradual traffic shifting can be achieved by updating the endpoint's variant weights over time.

Answer

Create a CloudWatch alarm to automatically rollback

Answer

Enable data capture on the endpoint

Answer

Use SageMaker Model Registry to approve the new model

Question 15

A company uses Amazon SageMaker to host a real-time inference endpoint for a fraud detection model. The endpoint is deployed with three instances of ml.m5.large. The model processes each request in about 200 ms. Lately, users report occasional timeouts (requests taking >5 seconds). The team suspects model drift or data skew. What is the MOST likely cause and solution?

Accepted Answer

A recent change increased the average input size, causing longer inference time; investigate input preprocessing.. The symptom of occasional timeouts (>5 seconds) on a model that normally processes requests in ~200 ms suggests that a recent change in input data characteristics (e.g., larger payloads or more complex features) is causing sporadic latency spikes. Investigating input preprocessing can identify if data skew or increased input size is overwhelming the model's inference path, which is a common monitoring concern in SageMaker real-time endpoints.

Answer

The instances are under-provisioned; switch to ml.m5.xlarge instances.

Answer

The endpoint is experiencing too many concurrent requests; add more instances.

Answer

Model drift caused the model to become computationally heavier; retrain the model.