MLS-C01 practice questions 251–260

1. Question 251 · 368 in bank

   Select 2

   A company wants to detect credit card fraud. The company has observed that an average of 2% of credit card transactions are fraudulent. A data scientist trains a classifier on a year's worth of credit card transaction data. The classifier needs to identify the fraudulent transactions. The company wants to accurately capture as many fraudulent transactions as possible. Which metrics should the data scientist use to optimize the classifier? (Choose two.)

   • ASpecificity
   • BFalse positive rate
   • CAccuracy
   • DF1 score
   • ETrue positive rate

   Show answer

   Answer: DE

   Explanation

   D. F1 score and E. True positive rate (Recall). F1 Score: This is the harmonic mean of precision and recall, providing a balance between the two. It is particularly useful when you need to balance the importance of both false positives and false negatives. True Positive Rate (Recall): This measures the proportion of actual fraudulent transactions that are correctly identified. High recall ensures that most fraudulent transactions are detected.

   Source: Adapted from community discussion — MultiCloudIronMan

   Next question

2. Question 252 · 368 in bank

   A data scientist is designing a repository that will contain many images of vehicles. The repository must scale automatically in size to store new images every day. The repository must support versioning of the images. The data scientist must implement a solution that maintains multiple immediately accessible copies of the data in different AWS Regions. Which solution will meet these requirements?

   • AAmazon S3 with S3 Cross-Region Replication (CRR)
   • BAmazon Elastic Block Store (Amazon EBS) with snapshots that are shared in a secondary Region
   • CAmazon Elastic File System (Amazon EFS) Standard storage that is configured with Regional availability
   • DAWS Storage Gateway Volume Gateway

   Show answer

   Answer: A

   Explanation

   S3: scalable versioned object storage CRR - Automatically replicates objects from source to destination buckets in a different AWS Regions

   Source: Adapted from community discussion — AIWave

   References

   • AWS documentation

   Next question

3. Question 253 · 368 in bank

   An ecommerce company wants to update a production real-time machine learning (ML) recommendation engine API that uses Amazon SageMaker. The company wants to release a new model but does not want to make changes to applications that rely on the API. The company also wants to evaluate the performance of the new model in production traffic before the company fully rolls out the new model to all users. Which solution will meet these requirements with the LEAST operational overhead?

   • ACreate a new SageMaker endpoint for the new model. Configure an Application Load Balancer (ALB) to distribute traffic between the old model and the new model.
   • BModify the existing endpoint to use SageMaker production variants to distribute traffic between the old model and the new model.
   • CModify the existing endpoint to use SageMaker batch transform to distribute traffic between the old model and the new model.
   • DCreate a new SageMaker endpoint for the new model. Configure a Network Load Balancer (NLB) to distribute traffic between the old model and the new model.

   Show answer

   Answer: B

   References

   • AWS documentation

   Next question

4. Question 254 · 368 in bank

   A machine learning (ML) specialist at a manufacturing company uses Amazon SageMaker DeepAR to forecast input materials and energy requirements for the company. Most of the data in the training dataset is missing values for the target variable. The company stores the training dataset as JSON files. The ML specialist develop a solution by using Amazon SageMaker DeepAR to account for the missing values in the training dataset. Which approach will meet these requirements with the LEAST development effort?

   • AImpute the missing values by using the linear regression method. Use the entire dataset and the imputed values to train the DeepAR model.
   • BReplace the missing values with not a number (NaN). Use the entire dataset and the encoded missing values to train the DeepAR model.
   • CImpute the missing values by using a forward fill. Use the entire dataset and the imputed values to train the DeepAR model.
   • DImpute the missing values by using the mean value. Use the entire dataset and the imputed values to train the DeepAR model.

   Show answer

   Answer: B

   Explanation

   DeepAR is a supervised RNNs and is able to handle missing values directly within the model. Instead of pre-processing the data to impute missing values externally, DeepAR can work directly with missing values encoded as NaN.

   Source: Adapted from community discussion — AIWave

   References

   • AWS documentation

   Next question

5. Question 255 · 368 in bank

   A law firm handles thousands of contracts every day. Every contract must be signed. Currently, a lawyer manually checks all contracts for signatures. The law firm is developing a machine learning (ML) solution to automate signature detection for each contract. The ML solution must also provide a confidence score for each contract page. Which Amazon Textract API action can the law firm use to generate a confidence score for each page of each contract?

   • AUse the AnalyzeDocument API action. Set the FeatureTypes parameter to SIGNATURES. Return the confidence scores for each page.
   • BUse the Prediction API call on the documents. Return the signatures and confidence scores for each page.
   • CUse the StartDocumentAnalysis API action to detect the signatures. Return the confidence scores for each page.
   • DUse the GetDocumentAnalysis API action to detect the signatures. Return the confidence scores for each page.

   Show answer

   Answer: A

   Explanation

   A. The AnalyzeDocument API action in Amazon Textract allows for the analysis of various features within a document, including signatures. By setting the FeatureTypes parameter to SIGNATURES, the law firm can instruct Textract to specifically focus on detecting signatures within the contracts. Additionally, Textract provides confidence scores for detected elements, including signatures. Therefore, by using this action and specifying the FeatureTypes parameter, the law firm can receive confidence scores for each page of each contract, facilitating the automation of signature detection.

   Source: Adapted from community discussion — ggrodskiy

   References

   • AWS documentation

   Next question

6. Question 256 · 368 in bank

   Select 2

   A gaming company has launched an online game where people can start playing for free, but they need to pay if they choose to use certain features. The company needs to build an automated system to predict whether or not a new user will become a paid user within 1 year. The company has gathered a labeled dataset from 1 million users. The training dataset consists of 1,000 positive samples (from users who ended up paying within 1 year) and 999,000 negative samples (from users who did not use any paid features). Each data sample consists of 200 features including user age, device, location, and play patterns. Using this dataset for training, the Data Science team trained a random forest model that converged with over 99% accuracy on the training set. However, the prediction results on a test dataset were not satisfactory Which of the following approaches should the Data Science team take to mitigate this issue? (Choose two.)

   • AAdd more deep trees to the random forest to enable the model to learn more features.
   • BInclude a copy of the samples in the test dataset in the training dataset.
   • CGenerate more positive samples by duplicating the positive samples and adding a small amount of noise to the duplicated data.
   • DChange the cost function so that false negatives have a higher impact on the cost value than false positives.
   • EChange the cost function so that false positives have a higher impact on the cost value than false negatives.

   Show answer

   Answer: CD

   Explanation

   I think it should be CD C: because we need a balance dataset D: The number of positive samples is large so model tends to predict 0 (negative) for all cases leading to False Negative problem. We should minimize that. My opinion

   Source: Adapted from community discussion — Phong

   Next question

7. Question 257 · 368 in bank

   Select 3

   A company that operates oil platforms uses drones to photograph locations on oil platforms that are difficult for humans to access to search for corrosion. Experienced engineers review the photos to determine the severity of corrosion. There can be several corroded areas in a single photo. The engineers determine whether the identified corrosion needs to be fixed immediately, scheduled for future maintenance, or requires no action. The corrosion appears in an average of 0.1% of all photos. A data science team needs to create a solution that automates the process of reviewing the photos and classifying the need for maintenance. Which combination of steps will meet these requirements? (Choose three.)

   • AUse an object detection algorithm to train a model to identify corrosion areas of a photo.
   • BUse Amazon Rekognition with label detection on the photos.
   • CUse a k-means clustering algorithm to train a model to classify the severity of corrosion in a photo.
   • DUse an XGBoost algorithm to train a model to classify the severity of corrosion in a photo.
   • EPerform image augmentation on photos that contain corrosion.
   • FPerform image augmentation on photos that do not contain corrosion.

   Show answer

   Answer: ADE

   Explanation

   A. Use an object detection algorithm: This can help identify corrosion areas in a photo. E. Perform image augmentation on photos with corrosion: This can improve the model’s ability to generalize by increasing the diversity of the training data. D. Use an XGBoost algorithm: This can classify the severity of the corrosion after the areas of corrosion have been identified. It’s effective for multi-class classification problems.

   Source: Adapted from community discussion — butaman

   References

   • AWS documentation

   Next question

8. Question 258 · 368 in bank

   A company maintains a 2 TB dataset that contains information about customer behaviors. The company stores the dataset in Amazon S3. The company stores a trained model container in Amazon Elastic Container Registry (Amazon ECR). A machine learning (ML) specialist needs to score a batch model for the dataset to predict customer behavior. The ML specialist must select a scalable approach to score the model. Which solution will meet these requirements MOST cost-effectively?

   • AScore the model by using AWS Batch managed Amazon EC2 Reserved Instances. Create an Amazon EC2 instance store volume and mount it to the Reserved Instances.
   • BScore the model by using AWS Batch managed Amazon EC2 Spot Instances. Create an Amazon FSx for Lustre volume and mount it to the Spot Instances.
   • CScore the model by using an Amazon SageMaker notebook on Amazon EC2 Reserved Instances. Create an Amazon EBS volume and mount it to the Reserved Instances.
   • DScore the model by using Amazon SageMaker notebook on Amazon EC2 Spot Instances. Create an Amazon Elastic File System (Amazon EFS) file system and mount it to the Spot Instances.

   Show answer

   Answer: B

   Next question

9. Question 259 · 368 in bank

   A data scientist is implementing a deep learning neural network model for an object detection task on images. The data scientist wants to experiment with a large number of parallel hyperparameter tuning jobs to find hyperparameters that optimize compute time. The data scientist must ensure that jobs that underperform are stopped. The data scientist must allocate computational resources to well-performing hyperparameter configurations. The data scientist is using the hyperparameter tuning job to tune the stochastic gradient descent (SGD) learning rate, momentum, epoch, and mini-batch size. Which technique will meet these requirements with LEAST computational time?

   • AGrid search
   • BRandom search
   • CBayesian optimization
   • DHyperband

   Show answer

   Answer: D

   Explanation

   . Hyperband: This technique is a bandit-based approach that allocates resources efficiently by running multiple configurations in parallel with varying durations. It eliminates poorly performing configurations early and focuses resources on promising ones, making it ideal for minimizing compute time.

   Source: Adapted from community discussion — KAST_424

   Next question

10. Question 260 · 368 in bank

    An agriculture company wants to improve crop yield forecasting for the upcoming season by using crop yields from the last three seasons. The company wants to compare the performance of its new scikit-learn model to the benchmark. A data scientist needs to package the code into a container that computes both the new model forecast and the benchmark. The data scientist wants AWS to be responsible for the operational maintenance of the container. Which solution will meet these requirements?

    • APackage the code as the training script for an Amazon SageMaker scikit-learn container.
    • BPackage the code into a custom-built container. Push the container to Amazon Elastic Container Registry (Amazon ECR).
    • CPackage the code into a custom-built container. Push the container to AWS Fargate.
    • DPackage the code by extending an Amazon SageMaker scikit-learn container.

    Show answer

    Answer: D

    Explanation

    Amazon SageMaker provides built-in containers for common machine learning frameworks, including scikit-learn, which are designed to handle operational maintenance such as patching, scaling, and monitoring.https://docs.aws.amazon.com/sagemaker/latest/dg/pre-built-docker-containers-scikit-learn-spark.html

    Source: Adapted from community discussion — Tkhan1

    References

    • AWS documentation

    Next question

Keyboard:J Kmove between questionsA–Dselect an answerRshow or hide the answer← →change page

MLS-C01 practice, question 251. 368 questions available.