MLS-C01 Practice Test 5 — 25 Questions

Question 1

A data analyst is examining a scatter plot of two variables and notices a strong positive correlation. Which of the following is a valid conclusion?

Accepted Answer

The two variables are related, but causation cannot be inferred. A strong positive correlation indicates that the two variables are related, but correlation alone does not imply causation. Option A is incorrect because correlation does not necessarily imply a linear relationship; it could be non-linear or monotonic. Option B is incorrect because correlation does not imply causation. Option D is incorrect because correlation does not guarantee accurate prediction; prediction requires a well-fitted model and additional validation.

Answer

The relationship is linear

Answer

One variable causes the other

Answer

The relationship can be used to accurately predict one variable from the other

Question 2

Which TWO AWS services can be used to deploy a machine learning model for serverless inference? (Choose 2.)

Accepted Answer

Amazon SageMaker Serverless Inference. Amazon SageMaker Serverless Inference (A) is correct because it is a purpose-built SageMaker hosting option that automatically provisions and scales compute for model inference without managing servers, scaling to zero when there is no traffic, making it ideal for serverless ML inference. AWS Lambda (B) is also correct because you can package a trained model (or call one) inside a Lambda function and expose it via API Gateway or a function URL, running inference with no server management and paying only per invocation. Amazon EMR (C) is a big-data processing service for Hadoop/Spark clusters, not a serverless inference hosting option. Amazon ECS with Fargate (D) runs containers on serverless compute but still requires you to manage tasks, services, and scaling, so it is not a fully serverless inference service. AWS Batch (E) is for batch computing job scheduling, not for serving low-latency serverless inference requests.

Answer

Amazon EMR

Answer

Amazon ECS with Fargate

Answer

AWS Batch

Question 3

A data scientist is training a deep learning model on Amazon SageMaker for image classification. The training is taking a long time and the GPU utilization is consistently below 30%. What should the data scientist do to improve GPU utilization and reduce training time?

Accepted Answer

Increase the batch size.. Low GPU utilization (below 30%) indicates that the GPU is spending most of its time waiting for data to process, often due to small batch sizes that underutilize the GPU's parallel compute capacity. Increasing the batch size allows the GPU to process more samples per forward/backward pass, improving arithmetic intensity and hardware utilization, which directly reduces total training time on SageMaker.

Answer

Use early stopping to stop training earlier.

Answer

Switch to a CPU-only instance.

Answer

Reduce the number of layers in the model.

Question 4

A data scientist is analyzing a dataset with 100 features and wants to identify which features are most correlated with the target variable. Which AWS service is most appropriate for this task?

Accepted Answer

Amazon SageMaker Data Wrangler. Amazon SageMaker Data Wrangler is purpose-built for data preparation and exploratory analysis, including correlation matrices, feature-target correlation, and feature importance visualizations. It integrates directly with SageMaker Studio and can export the analysis into a processing job or pipeline, making it the most appropriate service for identifying features correlated with a target variable.

Answer

Amazon QuickSight

Answer

Amazon Athena

Answer

AWS Glue DataBrew

Question 5

A company has deployed a machine learning model on Amazon SageMaker for real-time inference. The endpoint uses a single ml.c5.xlarge instance. Recently, the traffic has increased, and the endpoint is returning HTTP 503 (Service Unavailable) errors during peak hours. The CloudWatch metrics show that the CPU utilization is consistently above 90% during peak times, and the Invocations metric shows that requests are being throttled. The data science team has already optimized the model to reduce inference time by 20%, but the errors persist. The company needs to resolve the issue without increasing costs significantly. Which course of action should be taken?

Accepted Answer

Configure auto-scaling for the endpoint to add instances based on CPU utilization. The endpoint is CPU-bound (consistently above 90%) and throttling requests, causing 503 errors. Auto-scaling based on CPU utilization adds more instances behind the endpoint, distributing load and resolving throttling without over-provisioning a larger instance. This is cost-effective because instances are added only during peak times and removed when demand drops.

Answer

Change the instance type to a larger size, such as ml.c5.2xlarge

Answer

Switch to batch transform to process requests in batches

Answer

Use spot instances to reduce costs and add more instances

Question 6

A data scientist is training a neural network for image classification. The training loss decreases but validation loss increases after a few epochs. Which TWO actions should be taken to address this?

Accepted Answer

Implement early stopping based on validation loss.. The scenario describes overfitting: training loss keeps falling while validation loss rises after a few epochs, meaning the model is memorizing training data rather than generalizing. Option A is correct because early stopping monitors validation loss and halts training at the point where validation performance stops improving, preventing the model from continuing to overfit. Option C is correct because increasing the dropout rate regularizes the network by randomly deactivating neurons during training, which reduces co-adaptation and improves generalization, directly countering the overfitting trend. Option B is not appropriate because raising the learning rate would likely destabilize training or worsen convergence, not fix overfitting. Option D is wrong because training for more epochs would let the model overfit even further as validation loss continues to climb. Option E is wrong because adding more convolutional layers increases model capacity, which typically exacerbates overfitting rather than resolving it.

Answer

Increase the learning rate.

Answer

Increase the number of epochs.

Answer

Add more convolutional layers.

Question 7

A data scientist is using Amazon SageMaker to train a model with a custom Docker container. The training script reads data from an S3 bucket and writes the model artifact to an S3 bucket. The training job fails with a 'NoSuchKey' error. What is the MOST likely cause?

Accepted Answer

The training data path specified in the input data channel is incorrect.. The 'NoSuchKey' error indicates that the specified key (file path) does not exist within the S3 bucket, which occurs when the training data path provided in the input data channel is incorrect. Option A is wrong because, while a script compatibility issue could cause errors, it would not manifest as a 'NoSuchKey' S3 error; it would likely result in a failure during container execution. Option C is wrong because if the Docker image were not available in Amazon ECR, SageMaker would fail with an 'ImageNotFoundException' or similar ECR-related error, not a 'NoSuchKey' S3 error. Option D is wrong because insufficient S3 permissions (e.g., missing s3:GetObject) would result in an 'AccessDenied' error, not 'NoSuchKey'.

Answer

The training script is not compatible with the Docker image.

Answer

The Docker image is not available in Amazon ECR.

Answer

The SageMaker execution role does not have s3:GetObject permission.

Question 8

A data scientist wants to understand the distribution of a categorical feature with 100 unique values. Which visualization is most appropriate?

Accepted Answer

Bar chart. A bar chart is the most appropriate visualization for displaying the distribution of a categorical feature with 100 unique values because it uses discrete bars to represent the frequency or proportion of each category. Unlike a histogram, which requires continuous numeric bins, a bar chart preserves the distinct categories and allows clear comparison of counts across all 100 levels.

Answer

Histogram

Answer

Scatter plot

Answer

Pie chart

Question 9

A company wants to use Amazon SageMaker to train a deep learning model using a custom TensorFlow script. The data is stored in an S3 bucket. Which SageMaker API operation should be used to launch the training job?

Accepted Answer

CreateTrainingJob. The correct API operation to launch a training job in Amazon SageMaker is CreateTrainingJob. This operation specifies the training algorithm (or custom script), resource configuration (instance type and count), input data configuration (pointing to S3), and output location for the model artifacts. It directly initiates the training process on SageMaker-managed infrastructure.

Answer

CreateHyperParameterTuningJob

Answer

CreateEndpoint

Answer

CreateTransformJob

Question 10

A machine learning engineer is analyzing a dataset that contains a categorical feature 'country' with 200 unique values. The target variable is binary. The engineer wants to use this feature in a linear model. Which encoding method should be applied during EDA to prepare the data for modeling, considering the high cardinality?

Accepted Answer

Target encoding with cross-validation. Target encoding with cross-validation is suitable for high-cardinality categorical features because it replaces each category with the mean of the target variable for that category, reducing dimensionality while capturing the relationship with the target. Cross-validation prevents data leakage and overfitting, which is crucial when the feature has many unique values like 200 countries.

Answer

Label encoding

Answer

Frequency encoding

Answer

One-hot encoding

Question 11

A machine learning engineer notices that the target variable in a regression dataset has a long-tailed distribution. Which visualization technique is most appropriate to assess the distribution before applying a log transformation?

Accepted Answer

Histogram with density curve. (Histogram with density curve) is the most appropriate visualization for assessing a long-tailed distribution because it clearly shows the shape, spread, and tail behavior of the target variable. A histogram with an overlaid density curve helps identify skewness and the need for a log transformation. Option A (Bar chart) is for categorical data, not continuous distributions. Option C (Box plot) provides quartiles and outliers but does not fully reveal the distribution shape, especially the length of tails. Option D (Scatter plot) visualizes relationships between two variables, not a univariate distribution.

Answer

Bar chart

Answer

Box plot

Answer

Scatter plot

Question 12

A data scientist needs to query a dataset stored as Parquet files in Amazon S3 using standard SQL without managing any infrastructure. Which service should they use?

Accepted Answer

Amazon Athena. Amazon Athena is a serverless interactive query service that allows you to query data stored in Amazon S3 using standard SQL without any infrastructure to manage. It directly reads Parquet files from S3 and executes queries using Presto under the hood, making it the ideal choice for this use case.

Answer

Amazon QuickSight

Answer

AWS Glue

Answer

Amazon Redshift

Question 13

A team is analyzing a dataset with many categorical features that have high cardinality (e.g., ZIP code, user ID). They want to explore relationships between these features and a continuous target variable. Which approach is most appropriate for visualizing these relationships without overwhelming the viewer?

Accepted Answer

Group categories into top K levels and use a box plot for each group.. Grouping high-cardinality categorical features into top K levels and using box plots for each group is the most appropriate approach. This reduces the number of categories to a manageable number, allowing clear visualization of the distribution of the continuous target variable across the most frequent categories. It avoids overwhelming the viewer with too many categories while still revealing relationships.

Answer

Compute a correlation matrix using Pearson correlation.

Answer

Create a scatter plot with each category as a different color.

Answer

Use a heatmap to show pairwise chi-square statistics.

Question 14

A data scientist is performing EDA on a dataset that contains customer demographics and purchase history. The dataset has a column 'age' with some values that are negative or unreasonably high (e.g., 200). The scientist wants to identify and handle these outliers. The scientist is using a SageMaker notebook with pandas. Which approach should the scientist take to effectively handle these outliers?

Accepted Answer

Define reasonable bounds based on domain knowledge and filter or cap the outliers. The most appropriate approach is to define reasonable bounds based on domain knowledge (e.g., 0-120) and filter out or cap the outliers. Option A is incorrect because standard scaling does not handle outliers; it will still be influenced by extreme values. Option B is incorrect because imputing with the mean can distort the distribution when outliers are present. Option D is incorrect because removing the entire column discards valuable information.

Answer

Apply standard scaling to the 'age' column

Answer

Impute the outlier values with the mean of the column

Answer

Remove the 'age' column entirely

Question 15

A company wants to use Amazon SageMaker to train a model on a dataset stored in Amazon S3. The dataset is 100 GB and consists of millions of small JSON files. What should the data engineering team do to optimize training performance?

Accepted Answer

Combine the small JSON files into larger Parquet files using a Spark job on Amazon EMR.. Combining millions of small JSON files into larger Parquet files using a Spark job on Amazon EMR is correct because it reduces the overhead of S3 LIST and GET requests during training. Parquet's columnar format also improves compression and allows SageMaker to read only the necessary columns, significantly accelerating I/O-bound training workloads.

Answer

Copy the data to an Amazon EBS volume attached to the training instance.

Answer

Use Amazon Athena to convert the data into a single CSV file.

Answer

Use S3 Select to filter data before training.

Question 16

A machine learning engineer is working on a regression problem to predict house prices. The dataset contains 500,000 rows and 20 features, including 'sqft_living', 'bedrooms', 'bathrooms', 'floors', 'waterfront', 'view', 'condition', 'grade', 'yr_built', 'zipcode', and 'lat'. After performing exploratory data analysis, the engineer notices that the 'sqft_living' feature has a right-skewed distribution with a long tail. The 'zipcode' feature is categorical with 70 unique values. The 'lat' feature is continuous. The engineer wants to prepare the data for a linear regression model. Which action should the engineer take to improve model performance?

Accepted Answer

Apply a log transformation to the 'sqft_living' feature.. Linear regression assumes that features are approximately normally distributed, and a right-skewed distribution like 'sqft_living' can violate this assumption, leading to poor model performance. Applying a log transformation compresses the long tail, making the distribution more symmetric and helping the model learn a linear relationship between the feature and the target. This is a standard preprocessing step for skewed features in regression tasks.

Answer

Remove the 'sqft_living' feature because it violates the normality assumption.

Answer

One-hot encode the 'zipcode' feature to capture location effects.

Answer

Apply standard scaling (z-score) to the 'sqft_living' feature.

Question 17

A data scientist is using Amazon SageMaker to train a linear regression model. The dataset has outliers. Which TWO techniques can help reduce the impact of outliers? (Choose TWO.)

Accepted Answer

Trim the dataset to remove extreme values. Option A is correct because trimming (or winsorizing) the dataset removes or caps extreme values, so the linear regression fit is no longer dominated by outliers that would otherwise pull the coefficients disproportionately. Option D is correct because Huber loss behaves quadratically for small residuals but linearly for large residuals, so outliers contribute bounded gradients and have far less influence on the learned coefficients than with squared error. Option B is not appropriate because adding features does not address the influence of outliers and can even worsen overfitting. Option C is not appropriate because L1 regularization penalizes coefficient magnitude to induce sparsity, not to down-weight outlier residuals. Option E is not appropriate because standardization only rescales features to zero mean and unit variance; it does not reduce the leverage or residual impact of outliers.

Answer

Add more features

Answer

Apply L1 regularization

Answer

Standardize the features

Question 18

A team is training a deep learning model on Amazon SageMaker using a large dataset stored in S3. The training job is taking a long time, and the team suspects that data loading is the bottleneck. The dataset consists of many small files (average size 10KB). Which change would most effectively reduce the I/O bottleneck?

Accepted Answer

Combine the small files into larger files (e.g., TFRecord format). Combining small files into larger ones (e.g., TFRecord or Parquet) reduces the number of S3 GET requests and improves throughput. Using Pipe mode reads data sequentially, which is less efficient for random access. Increasing instance count or using P3 instances addresses compute, not I/O. Amazon EFS is not recommended for training jobs due to higher latency.

Answer

Use SageMaker Pipe mode instead of File mode

Answer

Increase the number of training instances

Answer

Use a P3 instance type for better GPU performance

Question 19

Which TWO metrics are suitable for evaluating a regression model? (Select TWO.)

Accepted Answer

Root Mean Squared Error (RMSE). Root Mean Squared Error (RMSE) (B) is correct because it is a standard regression metric that measures the average magnitude of prediction errors by taking the square root of the mean of squared differences between predicted and actual continuous values, expressed in the same units as the target. R-squared (C) is also correct because it quantifies the proportion of variance in the dependent variable explained by the model, making it a common goodness-of-fit measure for regression. Accuracy (A) is not suitable because it is a classification metric that requires exact matches of discrete labels, which is meaningless for continuous predictions. F1-score (D) and Precision (E) are classification metrics derived from the confusion matrix (true/false positives and negatives) and do not apply to regression tasks.

Answer

Accuracy

Answer

F1-score

Answer

Precision

Question 20

A data scientist is training a binary classification model on an imbalanced dataset where the positive class represents only 1% of the data. The model achieves 99% accuracy but fails to identify most positive cases. Which metric should the data scientist use to evaluate model performance?

Accepted Answer

F1 score. The F1 score is the harmonic mean of precision and recall, making it ideal for imbalanced datasets where accuracy is misleading. Since the model achieves 99% accuracy by simply predicting the majority class (negative), it fails to capture positive cases; F1 score penalizes this by balancing false positives and false negatives, providing a more truthful performance measure.

Answer

R-squared

Answer

Accuracy

Answer

RMSE