PMLE Architecting low-code ML solutions — All Questions With Answers

Refer to the exhibit.

```
# BigQuery ML model creation
CREATE OR REPLACE MODEL `mydataset.churn_model`
OPTIONS
  ( model_type='LOGISTIC_REG',
    auto_class_weights=TRUE,
    input_label_cols=['churned'] )
AS
SELECT
  * EXCEPT(customer_id, churn_date)
FROM `mydataset.training_data`
WHERE churn_date IS NOT NULL;

# Evaluation query
SELECT * FROM ML.EVALUATE(MODEL `mydataset.churn_model`);

# Prediction query
SELECT * FROM ML.PREDICT(MODEL `mydataset.churn_model`,
  TABLE `mydataset.new_customers`);
```

Refer to the exhibit:
$ gcloud ai endpoints deploy-model $ENDPOINT_ID \
  --model $MODEL_ID \
  --display-name=my-model \
  --machine-type=n1-standard-2 \
  --min-replica-count=1 \
  --max-replica-count=5 \
  --traffic-split=0=100

ERROR: (gcloud.ai.endpoints.deploy-model) RESOURCE_EXHAUSTED: The machine type n1-standard-2 is not available in region us-central1 for AutoML models.

Refer to the exhibit:
{
  "textPayload": "Prediction failed: Model not ready or not deployed.",
  "resource": {
    "type": "ai_platform_endpoint",
    "labels": {
      "endpoint_id": "12345678",
      "model_id": "87654321"
    }
  },
  "severity": "ERROR"
}

bq query --use_legacy_sql=false 'SELECT * FROM ML.PREDICT(MODEL mydataset.mymodel, (SELECT * FROM mydataset.newdata))'

{
  "name": "projects/my-project/locations/us-central1/endpoints/my-endpoint",
  "displayName": "my-endpoint",
  "deployedModels": [
    {
      "model": "projects/my-project/locations/us-central1/models/12345",
      "displayName": "mymodel",
      "autoscalingMetricSpecs": [
        {
          "metricName": "aiplatform.googleapis.com/prediction/online/requests",
          "target": 100
        }
      ]
    }
  ],
  "trafficSplit": {
    "12345": 100
  }
}

Refer to the exhibit.

```
CREATE OR REPLACE MODEL `mydataset.housing_model`
OPTIONS
  (model_type='linear_reg',
   input_label_cols=['price'],
   data_split_method='custom',
   data_split_col='split_flag')
AS
SELECT * FROM `mydataset.housing_data`
```

The table `housing_data` has 1000 rows. The `split_flag` column contains only NULL values. The model creation fails with the error: "Invalid state: The number of training data rows is 0."

CREATE OR REPLACE MODEL `mydataset.my_model`
TRANSFORM(
  feature1,
  feature2,
  ML.IMPUTER(feature3) OVER (feature1) AS feature3_imputed,
  ML.STANDARD_SCALER(feature4) OVER () AS feature4_scaled
)
OPTIONS(
  model_type='logistic_reg',
  input_label_cols=['label']
)
AS
SELECT * FROM `mydataset.mytable`