Drawing business conclusions

In this section, we'll use our ML model, and we'll understand how many times the BigQuery ML model is able to predict the actual outcome.

Using the default threshold of 0.5, let's see how many times the ML model is able to correctly identify when a driver will get a tip:

SELECT COUNT(*)
FROM (
      SELECT predicted_will_get_tip, predicted_will_get_tip_probs, will_get_tip actual_tip
      FROM
        ML.PREDICT(MODEL`05_chicago_taxi.binary_classification_version_5`,
          (
            SELECT
              trip_seconds,
              fare,
        ...