ML model training in Snowflake - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to create a machine learning model in Snowflake.

Snowflake

CREATE MODEL my_model FROM my_table PREDICTING target USING [1]; 

Drag options to blanks, or click blank then click option'

ASELECT

BINSERT

CLINEAR_REGRESSION

DUPDATE

Attempts:

3 left

💡 Hint

Common Mistakes

Using SQL commands like SELECT or INSERT instead of a model type.

Omitting the model type after PREDICTING.

The CREATE MODEL statement requires specifying the model type, such as LINEAR_REGRESSION, to train a model in Snowflake.

2fill in blank

medium

Complete the code to specify the input features for the model training.

Snowflake

CREATE MODEL my_model FROM my_table PREDICTING target USING LINEAR_REGRESSION INPUT ([1]);

Drag options to blanks, or click blank then click option'

AWHERE feature1 > 0

Btarget, feature1

CSELECT *

Dfeature1, feature2, feature3

Attempts:

3 left

💡 Hint

Common Mistakes

Including the target column as a feature.

Using SQL clauses like SELECT or WHERE inside INPUT.

The INPUT clause lists the columns used as input features for training the model. Only feature columns should be included, not the target.

3fill in blank

hard

Fix the error in the model training command by completing the missing clause.

Snowflake

CREATE MODEL my_model [1] my_table PREDICTING target USING LINEAR_REGRESSION INPUT (feature1, feature2);

Drag options to blanks, or click blank then click option'

AFROM

BWHERE

CSELECT

DJOIN

Attempts:

3 left

💡 Hint

Common Mistakes

Using WHERE or JOIN instead of FROM after INPUT.

Omitting the FROM clause entirely.

The FROM clause specifies the source table for training data in the model creation statement.

4fill in blank

hard

Fill both blanks to evaluate the trained model using SQL.

Snowflake

SELECT predicted_label, actual_label FROM TABLE([1](MODEL => 'my_model', DATA => [2]));

Drag options to blanks, or click blank then click option'

AML.PREDICT

Bmy_table

Ctraining_data

DML.TRAIN

Attempts:

3 left

💡 Hint

Common Mistakes

Using ML.TRAIN instead of ML.PREDICT for predictions.

Using a wrong table name or a keyword instead of a table.

The ML.PREDICT function is used to get predictions from the model, and the data table (e.g., my_table) provides the input data for prediction.

5fill in blank

hard

Fill all three blanks to create a model, train it, and evaluate accuracy.

Snowflake

CREATE MODEL my_model FROM training_data PREDICTING target USING [1] INPUT (feature1, feature2);

SELECT accuracy FROM TABLE([2](MODEL => 'my_model', DATA => test_data)) WHERE metric = '[3]';

Drag options to blanks, or click blank then click option'

ALOGISTIC_REGRESSION

BML.EVALUATE

Caccuracy

DML.PREDICT

Attempts:

3 left

💡 Hint

Common Mistakes

Using ML.PREDICT instead of ML.EVALUATE to get accuracy.

Using regression model type for classification tasks.

Misspelling the metric name.

First, specify the model type LOGISTIC_REGRESSION to create the model. Then use ML.EVALUATE to get evaluation metrics. Finally, filter by the metric name accuracy to see the model's accuracy.