Snowflake DSA-C03 試験概要:

Snowflake DSA-C03 試験シラバストピック:

Snowflake SnowPro Advanced: Data Scientist Certification 認定 DSA-C03 試験問題：

1. You're building a model to predict whether a user will click on an ad (binary classification: click or no-click) using Snowflake. The data is structured and includes features like user demographics, ad characteristics, and past user interactions. You've trained a logistic regression model using SNOWFLAKE.ML and are now evaluating its performance. You notice that while the overall accuracy is high (around 95%), the model performs poorly at predicting clicks (low recall for the 'click' class). Which of the following steps could you take to diagnose the issue and improve the model's ability to predict clicks, and how would you implement them using Snowflake SQL? SELECT ALL THAT APPLY.

A) Generate a confusion matrix using SQL to visualize the model's performance across both classes. Example SQL:

B) Calculate precision, recall, F I-score, and AUC for the 'click' class using SQL queries to get a more detailed understanding of the model's performance on the minority class. Example:

C) Reduce the amount of training data to avoid overfitting. Overfitting is known to produce low recall for the 'click' class.
D) Increase the complexity of the model by switching to a non-linear algorithm like Random Forest or Gradient Boosting without performing hyperparameter tuning, as more complex models always perform better.
E) Implement feature engineering by creating interaction terms or polynomial features from existing features using SQL, to capture potentially non-linear relationships between features and the target variable. Example:

2. A marketing team uses Snowflake to store customer purchase data'. They want to segment customers based on their spending habits using a derived feature called The 'PURCHASES' table has columns 'customer id' (IN T), 'purchase_date' (DATE), and 'purchase_amount' (NUMBER). The team needs a way to handle situations where a customer might have missing months (no purchases in a particular month). They want to impute a 0 spend for those months before calculating the average. Which approach provides the most accurate and robust calculation, especially when considering users with sparse purchase history?

A) Calculate the total spend for each customer and divide by the number of months since their first purchase: / DATEDlFF(month, CURRENT DATE()) GROUP BY customer_id'.
B) Create a view containing all months for each customer, left join with the 'PURCHASES' table, impute 0 for null 'purchase_amounts values, and then calculate the average spend. Requires creating a helper table for all the month.
C) Calculate the average spend only for customers with purchases in every month of the year. Ignore other customers in the analysis.
D) Use a window function to calculate the average spend over a fixed window of the last 3 months, ignoring missing months in the calculation.
E) Calculate the average monthly spend directly from the 'PURCHASES' table without accounting for missing months: 'AVG(purchase_amount) GROUP BY customer_id, date_trunc('month',

3. You are building a machine learning pipeline in Snowflake using Snowpark Python. You have completed the data preparation and feature engineering steps and now need to train a model. You want to track the performance of different model versions and hyperparameters using MLflow. You are considering these deployment strategies. Which of the deployment strategies allows automatic logging of metrics, parameters, and model artifacts to MLflow for each training run without requiring explicit MLflow logging code?

A) Train the model locally on your development machine and manually log metrics and artifacts to MLflow using the MLflow API. Then, deploy the trained model to Snowflake as a UDF or stored procedure.
B) Train the model within a Snowpark Python UDF. Use a Snowflake stage to store MLflow artifacts.
C) Use the Snowpark MLAPI and its integration with MLflow's autologging feature. Enable autologging before starting the training run. Deploy the model to Snowflake as a UDF.
D) Train the model within a Snowpark Python stored procedure. Use a Snowflake stage to store MLflow artifacts.
E) Train the model using Snowpark's DataFrame API directly in a Snowflake worksheet. Manually create a log file with metrics and model parameters and upload it to a Snowflake stage.

4. A data science team is using Snowpark ML to train a classification model. They want to log model metadata (e.g., training parameters, evaluation metrics) and artifacts (e.g., the serialized model file) for reproducibility and model governance purposes. Which of the following approaches is the most appropriate for integrating model logging and artifact management within the Snowpark ML workflow, minimizing operational overhead?

A) Use a custom Python function to manually write model metadata to a Snowflake table and store the model file in a Snowflake stage.
B) Employ a separate, external model management platform (e.g., Databricks MLflow, SageMaker Model Registry) and configure Snowpark to interact with it via API calls during model training and deployment.
C) Serialize the model object to a string and store it as a VARIANT column in a Snowflake table, alongside the model metadata.
D) Leverage the MLflow integration within Snowpark, utilizing its ability to track experiments, log parameters and metrics, and store model artifacts directly within Snowflake stages or external storage.
E) Only track basic model performance metrics in a Snowflake table and rely on code versioning (e.g., Git) for model artifact management.

5. You are building a machine learning pipeline that uses data stored in Snowflake. You want to connect a Jupyter Notebook running on your local machine to Snowflake using Snowpark. You need to securely authenticate to Snowflake and ensure that you are using a dedicated compute resource for your Snowpark session. Which of the following approaches is the MOST secure and efficient way to achieve this?

A) Use key pair authentication to connect to Snowflake, storing the private key securely on your local machine. Specify a dedicated virtual warehouse during session creation.
B) Configure OAuth authentication for your Snowflake account and use the OAuth token to establish a Snowpark session with a dedicated virtual warehouse.
C) Use the Snowflake Python connector with username and password and execute SQL commands to create a Snowpark DataFrame.
D) Hardcode a role with 'ACCOUNTADMIN' privileges in your Jupyter Notebook using username and password.
E) Store your Snowflake username and password directly in the Jupyter Notebook and create a Snowpark session using these credentials and the default Snowflake warehouse.

質問と回答：