ML model training in Snowflake - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

Output of Snowflake ML model training command

What is the output of the following Snowflake SQL command that trains a linear regression model?

Snowflake

CREATE OR REPLACE MODEL my_linear_model
  OPTIONS (model_type = 'linear_regression', input_label_cols = ('price')) AS
  SELECT sqft, bedrooms, bathrooms, price FROM housing_data;

A Model my_linear_model created successfully with linear_regression type.

B Model training failed due to missing input_label_cols option.

C Error: Model type linear_regression not supported in Snowflake.

D Syntax error: Missing FROM clause in SELECT statement.

💡 Hint

Check the syntax for CREATE MODEL in Snowflake and required options.

The command correctly creates a linear regression model named my_linear_model using the specified columns. Snowflake supports linear_regression as a model_type and requires input_label_cols to specify the target column.

❓ Model Choice

intermediate

Choosing the correct model type for classification in Snowflake

You want to train a model in Snowflake to predict if a customer will churn (yes/no). Which model_type option should you use in the CREATE MODEL statement?

A 'time_series_forecast'

B 'kmeans'

C 'linear_regression'

D 'logistic_regression'

💡 Hint

Classification problems with yes/no targets usually use logistic regression.

Logistic regression is used for binary classification tasks like predicting churn (yes/no). Linear regression is for continuous targets, kmeans is for clustering, and time_series_forecast is for forecasting time-based data.

❓ Hyperparameter

advanced

Effect of changing max_iterations in Snowflake model training

In Snowflake's CREATE MODEL statement, what is the effect of increasing the max_iterations option for a logistic regression model?

A It decreases the training time by limiting the number of features used.

B It increases the number of training cycles, potentially improving model accuracy but increasing training time.

C It changes the learning rate to speed up convergence.

D It automatically selects the best regularization parameter.

💡 Hint

max_iterations controls how many times the algorithm runs through the data.

max_iterations sets the maximum number of optimization steps during training. Increasing it allows the model more chances to improve but takes longer.

❓ Metrics

advanced

Interpreting model evaluation metrics in Snowflake

After training a binary classification model in Snowflake, you get these metrics: accuracy = 0.85, precision = 0.75, recall = 0.60. What does the recall value tell you about the model?

A The model correctly identifies 60% of all actual positive cases.

B The model correctly predicts 75% of its positive predictions.

C The model is 85% accurate overall on the test data.

D The model has 60% accuracy on negative cases.

💡 Hint

Recall measures how many actual positives were found.

Recall is the ratio of true positives found over all actual positives. A recall of 0.60 means the model finds 60% of the positive cases.

🔧 Debug

expert

Debugging Snowflake model training failure due to data type mismatch

You run this command to train a model but get an error: 'Data type mismatch for input_label_cols'. What is the likely cause? CREATE OR REPLACE MODEL churn_model OPTIONS (model_type = 'logistic_regression', input_label_cols = ('churn')) AS SELECT customer_id, churn FROM customer_data;

A The input_label_cols option should be a list of multiple columns.

B The model_type 'logistic_regression' is not supported for this data.

C The 'churn' column is not numeric or boolean, but a string type.

D The SELECT statement is missing required feature columns besides customer_id and churn.

💡 Hint

Check the data type of the label column used for training.

Snowflake requires the label column for classification to be numeric or boolean. If 'churn' is stored as text (e.g., 'yes'/'no'), it causes a data type mismatch error.