> ## Documentation Index
> Fetch the complete documentation index at: https://docs.driftodata.com/llms.txt
> Use this file to discover all available pages before exploring further.

# CREATE MODEL

`CREATE MODEL`: This statement allows you to invoke Drifto’s autoML
capabilities. `CREATE MODEL` creates a new model that is stored in your model
registry with the given name. By default, Drifto uses an internal
[MLflow](www.mlflow.org) instance, but can be configured to use an external
instance as well. Like regular create statements, one should use
`CREATE OR REPLACE MODEL` to overwrite an existing model with the same name.
You may also use this statement to copy a pre-existing ML model (see examples below).

The `SELECT` subquery in the create statement should return the full training
dataset (held-out sets are automatically subsampled but can be manually specified
as well).

Each row in the dataset corresponds to one data point for ML training and
each column is a feature. The `WITH` clause is optional but for ML training
and each column is a feature. The `WITH` clause is optional but the target
column to be called `target`.

The ML training routine will train and validate a variety of ML architectures.
You can control this automated model search with argument flags.

### Syntax

```sql
CREATE MODEL <MODEL> WITH (**kwargs) AS <SELECT>;
```

### Keyword Args

<ParamField query="target" type="column or string literal" default="target">
  The name of the target column.
</ParamField>

<ParamField query="algo" type="string literal" default="auto">
  The training algorithm.
  Options = `{'auto', 'gbm', 'linear', 'dnn', 'drf'}`.

  * The `'auto'` option selects an autoML routine that will try multiple
    different algorithms and hyperparameters and select the best one.
  * The `'gbm'` option selects a gradient-boosting machine (GBM) algorithm.
  * The `'linear'` option selects a generalized linear model (GLM).
  * The `'dnn'` option selects a fully-connected feed-forward deep neural net.
  * The `'drf'` option selects a distributed random forest (DRF) architecture
    that includes extreme random forests (XRF).
</ParamField>

<ParamField query="ensemble" type="boolean (literal)" default="false">
  Allow autoML algorithm to generate model ensembles.
</ParamField>

<ParamField query="num_folds" type="int (literal)" default="0">
  The number of cross validation folds used in autoML. When set to 0,
  autoML does not use cross validation and instead will subsample a 10%
  validation and a 10% test set for validation and testing.
</ParamField>

<ParamField query="max_models" type="int (literal)" default="3">
  Maximum number of models to search for in an ML training routine.
</ParamField>

<ParamField query="seed" type="int (literal)" default="1">
  Random seed to use.
</ParamField>

<ParamField query="sort_metric" type="str (literal)" default="accuracy or mae">
  The figure-of-merit used to select the best model. The default is metric is
  accuracy for classification tasks and mae for regression tasks.
  Options (Classification) = `{'auc', 'accu', 'logloss'}`.
  Options (Regression) = `{'mse', 'mae', 'r2'}`.
</ParamField>

### Examples

```sql
CREATE MODEL my_model 
WITH (target=y, max_models=10, sort_metric='logloss') 
AS SELECT * FROM mydata;
```

```sql
-- This will overwrite an existing model.
CREATE MODEL OR REPLACE my_other_model 
WITH (max_models=1, algo='linear') 
AS SELECT * FROM mydata;
```

```sql
-- This will overwrite an existing model.
CREATE MODEL OR REPLACE my_other_model 
WITH (max_models=1, algo='linear') 
AS SELECT * FROM mydata;
```

```sql
-- This will create a copy of my_model.
-- Notice that here we've directly selected a model name.
CREATE MODEL my_model_copy 
AS SELECT my_model;
```