---
title: Evaluate Predictions from ML Models
sidebarTitle: Evaluate Predictions
---

## Description

The `EVALUATE` statement evaluates predictions based on [available metrics](https://scikit-learn.org/stable/modules/model_evaluation.html).

## Syntax

Here is the syntax:

```sql
EVALUATE metric_name
FROM (SELECT real_value AS actual,
             predicted_value AS prediction
      FROM table_name);
```

Where:

| Name              | Description                                                                                                                     |
| ----------------- | ------------------------------------------------------------------------------------------------------------------------------- |
| `metric_name`     | It is the name of the metric to be evaluated chosen from [here](https://scikit-learn.org/stable/modules/model_evaluation.html). |
| `real_value`      | It is the real value that will be compared with the predicted value.                                                            |
| `predicted_value` | It is the value predicted by the model.                                                                                         |
| `table_name`      | It is the table that stores corresponding real and predicted values.                                                            |

## Example

This example calculates the mean absolute error.

```sql
EVALUATE mean_absolute_error
FROM (SELECT column_name_that_stores_real_value AS actual,
             column_name_that_stores_predicted_value AS prediction
      FROM table);
```