MLS-C01 Practice Test 9 — 10 Questions

Question 1

A data scientist is performing exploratory data analysis on a dataset with mixed data types (numerical, categorical, text). The goal is to identify clusters of similar records. Which technique is most appropriate?

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

Accepted Answer

K-prototypes clustering. K-prototypes clustering is specifically designed to handle mixed data types (numerical and categorical) by combining K-means for numerical attributes and K-modes for categorical attributes. It uses a distance measure that integrates Euclidean distance for numerical features and Hamming distance for categorical features, making it suitable for datasets with mixed types. The other algorithms are not inherently designed for mixed data types without preprocessing.

Answer

DBSCAN

Answer

Hierarchical clustering

Answer

K-means clustering

Question 2

A data engineer needs to transform large CSV files stored in S3 into Parquet format and load them into a data warehouse for analysis. The transformation must be cost-effective and serverless. Which AWS service should be used?

Accepted Answer

AWS Glue. AWS Glue is the correct choice because it provides a fully managed, serverless ETL service that can automatically convert CSV files from S3 into Parquet format using its built-in Spark engine. It is cost-effective as you only pay for the resources consumed during the job execution, and it integrates directly with data warehouses like Amazon Redshift for loading transformed data.

Answer

Amazon Athena

Answer

Amazon EMR with Spark

Answer

AWS Data Pipeline

Question 3

A data scientist needs to understand the distribution of a numeric feature in a dataset stored in Amazon S3. Which AWS service can be used to run a quick exploratory query without setting up a server?

Accepted Answer

Amazon Athena. Amazon Athena is a serverless, interactive query service that allows you to analyze data directly in Amazon S3 using standard SQL. It requires no infrastructure setup or management, making it ideal for quick exploratory queries on data stored in S3. You simply point Athena to your S3 data, define a table in the AWS Glue Data Catalog, and run SQL queries to understand the distribution of a numeric feature. This aligns perfectly with the requirement to avoid server setup.

Answer

Amazon Redshift

Answer

Amazon EMR

Answer

AWS Glue

Question 4

A team is performing exploratory data analysis on a dataset containing 10 million records stored in Amazon S3. They want to sample the data efficiently to build a representative subset for initial modeling. Which sampling method should they use to minimize bias and ensure the sample reflects the population distribution?

Accepted Answer

Stratified random sampling. Stratified random sampling divides the population into homogeneous subgroups (strata) based on a relevant characteristic and then samples proportionally from each stratum. This ensures the sample mirrors the population's distribution across key dimensions, minimizing bias when the dataset has heterogeneous subgroups. For 10 million records with potentially skewed classes or categories, stratified sampling produces a more representative subset than simple random sampling.

Answer

Simple random sampling

Answer

Systematic sampling

Answer

Reservoir sampling

Question 5

A data pipeline uses AWS Glue to transform data from Amazon RDS to Amazon S3. The team wants to ensure that only new or updated records are processed in each run, minimizing cost and time. Which AWS Glue feature should be used?

Accepted Answer

Enable Glue Job Bookmarks.. Glue Job Bookmarks track processed data and persist state between job runs, enabling incremental processing of new or updated records from a source like Amazon RDS. This minimizes cost and time by avoiding full table scans and reprocessing unchanged data.

Answer

Use Glue triggers to run the job on a schedule.

Answer

Use Glue partition pruning to filter data.

Answer

Use Glue crawlers to detect new data.

Question 6

A data scientist is training a linear regression model on a dataset with 50 features. After training, they notice that the model performs well on training data but poorly on test data. They suspect overfitting. Which action should they take to reduce overfitting?

Accepted Answer

Add L2 regularization (Ridge regression). L2 regularization (Ridge regression) adds a penalty term proportional to the square of the magnitude of the coefficients to the loss function. This discourages the model from fitting the noise in the training data by shrinking the weights, which reduces variance and mitigates overfitting, improving generalization to the test data.

Answer

Use a larger learning rate

Answer

Add more features to the model

Answer

Increase the number of training epochs

Question 7

Which THREE factors should be considered when choosing between a parametric and a non-parametric machine learning model?

Accepted Answer

Non-parametric models are generally more flexible. Option A is correct because non-parametric models, such as k-NN, decision trees, or kernel SVMs, do not assume a fixed functional form and can adapt their complexity to the data, making them generally more flexible. Option C is correct because parametric models like linear or logistic regression estimate a fixed, small set of parameters, so training is typically faster than for non-parametric models that must retain or process much of the training data. Option E is correct because parametric models have strong structural assumptions and a fixed number of parameters, so they usually need less training data to estimate those parameters reliably. Option B is incorrect because non-parametric models typically have lower bias and higher variance, not the reverse. Option D is incorrect because non-parametric models are generally more prone to overfitting, not less, due to their higher flexibility.

Answer

Parametric models have lower bias than non-parametric models

Answer

Non-parametric models are less prone to overfitting

Question 8

A data scientist is training a deep learning model on Amazon SageMaker and wants to reduce the training time. Which TWO actions would help achieve this?

Accepted Answer

Use distributed training across multiple instances.. Option B is correct because distributed training across multiple instances (e.g., using SageMaker's distributed data parallel or model parallel libraries) splits the compute workload across nodes, directly reducing wall-clock training time for large deep learning models. Option D is correct because GPU-based instance types (e.g., ml.p3, ml.p4d, ml.g4dn) provide massively parallel floating-point throughput and high-bandwidth memory that accelerate the matrix operations dominating deep learning training compared to CPU instances. Option A is not correct because data augmentation increases the volume and variety of training data, which typically adds preprocessing and training overhead rather than reducing training time. Option C is not correct because SageMaker Automatic Model Tuning (hyperparameter tuning) runs many training jobs to find optimal hyperparameters, which increases overall time and cost rather than reducing a single training run's duration. Option E is not correct because SageMaker Managed Spot Training reduces cost by using spare EC2 capacity, not training time, and can even introduce interruptions and checkpoint-restart overhead.

Answer

Enable data augmentation.

Answer

Use SageMaker Automatic Model Tuning.

Answer

Use SageMaker Managed Spot Training.

Question 9

During EDA, a data scientist finds that two features have a Pearson correlation coefficient of 0.95. What is the primary concern when using these features together in a linear regression model?

Accepted Answer

Multicollinearity will make coefficient estimates unstable. A Pearson correlation coefficient of 0.95 indicates strong multicollinearity between the two features. Multicollinearity inflates the variance of coefficient estimates, making them unstable and difficult to interpret. Option A is wrong because redundant information leads to multicollinearity, not underfitting; underfitting occurs when the model is too simple. Option B is wrong because heteroscedasticity refers to non-constant variance of errors, not correlation between features. Option C is wrong because overfitting is more associated with model complexity and variance, not directly with redundant features; in fact, redundant features can cause numerical instability but not necessarily overfitting.

Answer

The model will underfit because of redundant information

Answer

Heteroscedasticity will be introduced

Answer

The model will overfit due to redundant features

Question 10

A data scientist is training a neural network on a dataset with 1 million images. The training loss decreases steadily but the validation loss starts to increase after 10 epochs. Which action should the scientist take to improve generalization?

Accepted Answer

Implement early stopping. Increasing validation loss while training loss decreases indicates overfitting. Early stopping (Option A) halts training when validation loss stops improving, directly preventing overfitting. Option B (adding more layers) increases model capacity and typically worsens overfitting. Option C (reducing learning rate) might slow training but does not directly stop overfitting. Option D (increasing epochs) would continue training and likely worsen overfitting.

Answer

Add more layers to the network

Answer

Reduce the learning rate

Answer

Increase the number of epochs