<note> <p> This operation creates a legacy predictor that does not include all the predictor functionalities provided by Amazon Forecast. To create a predictor that is compatible with all aspects of Forecast, use <a>CreateAutoPredictor</a>.</p> </note> <p>Creates an Amazon Forecast predictor.</p> <p>In the request, provide a dataset group and either specify an algorithm or let Amazon Forecast choose an algorithm for you using AutoML. If you specify an algorithm, you also can override algorithm-specific hyperparameters.</p> <p>Amazon Forecast uses the algorithm to train a predictor using the latest version of the datasets in the specified dataset group. You can then generate a forecast using the <a>CreateForecast</a> operation.</p> <p> To see the evaluation metrics, use the <a>GetAccuracyMetrics</a> operation. </p> <p>You can specify a featurization configuration to fill and aggregate the data fields in the <code>TARGET_TIME_SERIES</code> dataset to improve model training. For more information, see <a>FeaturizationConfig</a>.</p> <p>For RELATED_TIME_SERIES datasets, <code>CreatePredictor</code> verifies that the <code>DataFrequency</code> specified when the dataset was created matches the <code>ForecastFrequency</code>. TARGET_TIME_SERIES datasets don't have this restriction. Amazon Forecast also verifies the delimiter and timestamp format. For more information, see <a>howitworks-datasets-groups</a>.</p> <p>By default, predictors are trained and evaluated at the 0.1 (P10), 0.5 (P50), and 0.9 (P90) quantiles. You can choose custom forecast types to train and evaluate your predictor by setting the <code>ForecastTypes</code>. </p> <p> <b>AutoML</b> </p> <p>If you want Amazon Forecast to evaluate each algorithm and choose the one that minimizes the <code>objective function</code>, set <code>PerformAutoML</code> to <code>true</code>. The <code>objective function</code> is defined as the mean of the weighted losses over the forecast types. By default, these are the p10, p50, and p90 quantile losses. For more information, see <a>EvaluationResult</a>.</p> <p>When AutoML is enabled, the following properties are disallowed:</p> <ul> <li> <p> <code>AlgorithmArn</code> </p> </li> <li> <p> <code>HPOConfig</code> </p> </li> <li> <p> <code>PerformHPO</code> </p> </li> <li> <p> <code>TrainingParameters</code> </p> </li> </ul> <p>To get a list of all of your predictors, use the <a>ListPredictors</a> operation.</p> <note> <p>Before you can use the predictor to create a forecast, the <code>Status</code> of the predictor must be <code>ACTIVE</code>, signifying that training has completed. To get the status, use the <a>DescribePredictor</a> operation.</p> </note>

{
  "success": true,
  "data": {
    "id": "abc123",
    "created_at": "2025-01-01T00:00:00Z"
  }
}

{
  "success": false,
  "error": {
    "code": "VALIDATION_ERROR",
    "message": "Invalid request parameters"
  }
}

<note> <p> This operation creates a legacy predictor that does not include all the predictor functionalities provided by Amazon Forecast. To create a predictor that is compatible with all aspects of Forecast, use <a>CreateAutoPredictor</a>.</p> </note> <p>Creates an Amazon Forecast predictor.</p> <p>In the request, provide a dataset group and either specify an algorithm or let Amazon Forecast choose an algorithm for you using AutoML. If you specify an algorithm, you also can override algorithm-specific hyperparameters.</p> <p>Amazon Forecast uses the algorithm to train a predictor using the latest version of the datasets in the specified dataset group. You can then generate a forecast using the <a>CreateForecast</a> operation.</p> <p> To see the evaluation metrics, use the <a>GetAccuracyMetrics</a> operation. </p> <p>You can specify a featurization configuration to fill and aggregate the data fields in the <code>TARGET_TIME_SERIES</code> dataset to improve model training. For more information, see <a>FeaturizationConfig</a>.</p> <p>For RELATED_TIME_SERIES datasets, <code>CreatePredictor</code> verifies that the <code>DataFrequency</code> specified when the dataset was created matches the <code>ForecastFrequency</code>. TARGET_TIME_SERIES datasets don't have this restriction. Amazon Forecast also verifies the delimiter and timestamp format. For more information, see <a>howitworks-datasets-groups</a>.</p> <p>By default, predictors are trained and evaluated at the 0.1 (P10), 0.5 (P50), and 0.9 (P90) quantiles. You can choose custom forecast types to train and evaluate your predictor by setting the <code>ForecastTypes</code>. </p> <p> <b>AutoML</b> </p> <p>If you want Amazon Forecast to evaluate each algorithm and choose the one that minimizes the <code>objective function</code>, set <code>PerformAutoML</code> to <code>true</code>. The <code>objective function</code> is defined as the mean of the weighted losses over the forecast types. By default, these are the p10, p50, and p90 quantile losses. For more information, see <a>EvaluationResult</a>.</p> <p>When AutoML is enabled, the following properties are disallowed:</p> <ul> <li> <p> <code>AlgorithmArn</code> </p> </li> <li> <p> <code>HPOConfig</code> </p> </li> <li> <p> <code>PerformHPO</code> </p> </li> <li> <p> <code>TrainingParameters</code> </p> </li> </ul> <p>To get a list of all of your predictors, use the <a>ListPredictors</a> operation.</p> <note> <p>Before you can use the predictor to create a forecast, the <code>Status</code> of the predictor must be <code>ACTIVE</code>, signifying that training has completed. To get the status, use the <a>DescribePredictor</a> operation.</p> </note>

{
  "success": true,
  "data": {
    "id": "abc123",
    "created_at": "2025-01-01T00:00:00Z"
  }
}

{
  "success": false,
  "error": {
    "code": "VALIDATION_ERROR",
    "message": "Invalid request parameters"
  }
}