capabilities-forecast-predictions_intervals.mdx

---
title: "Prediction Intervals"
description: "Learn how to use the level parameter to generate prediction intervals that quantify forecast uncertainty."
icon: "chart-candlestick"
---

<CardGroup>
  <Card title="What are Prediction Intervals?" icon="circle-info">
    Prediction intervals measure the uncertainty around forecasted values. By specifying a confidence level, you can visualize the range in which future observations are expected to fall.
  </Card>
  <Card title="Key Parameter: level" icon="gear">
    The **level** parameter accepts values between 0 and 100 (including decimals). For example, `[80]` represents an 80% confidence interval.
  </Card>
</CardGroup>

## Overview

Use the `forecast` method's **level** parameter to generate prediction intervals. This helps quantify the uncertainty around your forecasts.

[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/Nixtla/nixtla/blob/main/nbs/docs/capabilities/forecast/10_prediction_intervals.ipynb)

<Steps>

  <Step title="Step 1: Import Dependencies">

  ```python Import Dependencies
  import pandas as pd
  from nixtla import NixtlaClient
  ```

  </Step>

  <Step title="Step 2: Initialize NixtlaClient">

  ```python Initialize NixtlaClient
  nixtla_client = NixtlaClient(
      # defaults to os.environ.get("NIXTLA_API_KEY")
      api_key='my_api_key_provided_by_nixtla'
  )
  ```

  </Step>

  <Step title="(Optional) Use an Azure AI Endpoint">
    <AccordionGroup>
      <Accordion title="Configuring an Azure AI Endpoint">
        <Check>
          **Use an Azure AI endpoint**
          To use an Azure AI endpoint, set the `base_url` argument as follows:
        </Check>
        ```python Azure AI Endpoint Configuration
        nixtla_client = NixtlaClient(
            base_url="your azure ai endpoint",
            api_key="your api_key"
        )
        ```
      </Accordion>
    </AccordionGroup>
  </Step>

  <Step title="Step 3: Load Dataset">

  ```python Load Dataset
  df = pd.read_csv(
      "https://raw.githubusercontent.com/Nixtla/transfer-learning-time-series/main/datasets/air_passengers.csv"
  )
  ```

  </Step>

  <Step title="Step 4: Generate Forecast with an 80% Interval">

  ```python Generate 80% Interval Forecast
  forecast_df = nixtla_client.forecast(
      df=df,
      h=12,
      time_col='timestamp',
      target_col='value',
      level=[80]
  )
  ```

  </Step>

  <Step title="Step 5: Plot Predictions and Intervals">

  ```python Plot Forecast and Intervals
  nixtla_client.plot(
      df=df,
      forecasts_df=forecast_df,
      time_col='timestamp',
      target_col='value',
      level=[80]
  )
  ```

  </Step>

</Steps>

<Info>
Logs indicate the validation and preprocessing steps, along with the inferred data frequency:
</Info>

<Accordion title="Forecasting Log Output">
```bash Log Output
INFO:nixtla.nixtla_client:Validating inputs...
INFO:nixtla.nixtla_client:Preprocessing dataframes...
INFO:nixtla.nixtla_client:Inferred freq: MS
INFO:nixtla.nixtla_client:Restricting input...
INFO:nixtla.nixtla_client:Calling Forecast Endpoint...
```
</Accordion>

<Frame caption="Forecast with an 80% Prediction Interval">
  ![Forecast](https://raw.githubusercontent.com/Nixtla/nixtla/readme_docs/nbs/_docs/docs/capabilities/forecast/10_prediction_intervals_files/figure-markdown_strict/cell-10-output-2.png)
</Frame>

<AccordionGroup>
  <Accordion title="Using Azure AI Models">
    <Info>
      **Available Models in Azure AI**
      If you are using an Azure AI endpoint, set the `model` parameter to `"azureai"`:
    </Info>
    ```python Azure AI Models
    nixtla_client.forecast(..., model="azureai")
    ```
    The public API supports two models:

      - `timegpt-1` (default)

      - `timegpt-1-long-horizon`



      See [this tutorial](https://docs.nixtla.io/docs/tutorials-long_horizon_forecasting) for guidance on using **timegpt-1-long-horizon**.

  </Accordion>
</AccordionGroup>

<Info>
For more information on uncertainty estimation, refer to the tutorials about [quantile forecasts](https://docs.nixtla.io/docs/tutorials-quantile_forecasts) and [prediction intervals](https://docs.nixtla.io/docs/tutorials-prediction_intervals).
</Info>