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

# Create Dynamic Column by Classification

> The **Classification** feature allows users to categorise dataset rows by applying labels based on text content from a selected column.

## **1. Select a Dataset**

Before setting up classification, ensure you have selected a dataset. If no dataset is available, follow the steps to **Add Dataset** on the Future AGI platform.

***

## **2. Accessing the Classification Interface**

To configure classification, navigate to your dataset and click the **+ Add Columns** button in the top-right menu. Scroll down to the **Dynamic Columns** section and select **Classification** to open the setup panel.

***

## **3. Configuring Classification Settings**

* **Name**: Assign a name to the new column where the classification results will be stored.
* **Column**: Select the dataset column that contains text data to be classified.
* **Labels**: Manually define classification labels by clicking **Add Label**. These labels should represent the possible categories for classification.
  * Example: If it is product reviews, you can set labels as "Positive", "Negative", and "Neutral".
* **Model**: Choose an AI model that will process the classification task.
* **Concurrency**: Define how many rows should be processed simultaneously for efficiency.

***

After configuring the settings, click **Test** to preview classification results on sample rows. If the classifications appear accurate, click **Create New Column** to apply classification across the dataset.

The new column will populate with predicted labels for each row based on the selected AI model.

***

## **Best Practices for Using Classification**

* **Ensure the selected column contains meaningful text data** for classification.
* **Define clear and distinct labels** to improve the accuracy of classification.
* **Adjust concurrency settings** based on dataset size for better processing efficiency.

***