# Label set editing and mapping ### Managing your Label Set You may wish to edit, export or delete labels within your Label Set. This feature has proved particularly useful for re-analysing photos when taxonomy changes occur! Managing your Label Set can be done within a Project under the **Label Sets** tab on the left-hand side menu in the ReefCloud Portal. 1. **View and Edit Label Set** via "Edit" under the "Actions" button. 2. You may also want to **Export the Label Set** as a .CSV file via the "Actions" button to share your unique Label Set with collaborators or for your records. 3. Label Sets can also be **Deleted** via the "Actions" button. {% hint style="danger" %} This should not be taken lightly, is computationally heavy taking some time to complete and is not reversable. Taking this action will also delete all point annotation data within a project associated with the original Label Set. {% endhint %}

A project Label Set can be easily managed via the Label Sets page.

4. **Filter** for Label Descriptions or Group Codes to find and edit a given label within the Label Set.

Edit the Label Set via the Actions button.

5. **Add New Labels** to the Label Set, possibly a category or taxon which you haven't come across before or didn't think to include in the original Label Set. For a new label, you will need to come up with a unique code that isn't the same as anything already used in your project. {% hint style="warning" %} There may be a risk for your model by having too many labels (e.g., 1000 labels). Overfitting and underfitting are two common problems in machine learning models. Overfitting occurs when a model learns the training data too well, including noise, and performs poorly on unseen data. Underfitting, on the other hand, happens when a model is too simple and fails to capture the underlying patterns in the data, resulting in poor performance on both training and unseen data. While "too many labels" might seem like it would only lead to overfitting, it can also contribute to underfitting if those labels are not actually informative or representative of the underlying data structure. {% endhint %} 6. **Sort Labels** using the column headings of the Label Set table. If needed, increase the number of labels viewed per page at the bottom of the scrollable box. 7. **Edit or Delete Existing Labels***.* Labels are editable apart from the unique code. {% hint style="danger" %} Deleting labels is a computationally heavy process and may take some time depending on the change and size of your project. It is also important to consider carefully whether deleting labels is necessary as it will have far reaching impacts across all points within your project regardless of whether the machine or human have applied the code. Reach out to the team for advice. {% endhint %} 8. Add a **Keyboard Shortcut Code** (as a numerical character) for labels of benthic taxa that appear frequently when annotating. Shortcut codes are optional. {% embed url="" %} Short video tutorial on how to manage and map a Label Set in the ReefCloud Portal. {% endembed %} ### Mapping your Label Set Label Set mapping is a crucial step that all ReefCloud users are encouraged to do. Mapping converts your imported labels into biological categories that are meaningful to ReefCloud, so your data can be graphed within your Project and compared between Projects (if you have opted to share your result summaries within the ReefCloud Public Dashboard). Visualisation of data within the Portal allows the user to [validate AI point annotations](/train-and-validate-ai/validated-ai-results.md) based on how well the machine has been trained. The human versus machine comparisons are displayed on the Portal Dashboard for hard coral, as well as the Model Report page. Project data is aggregated into 13 major benthic groups and displayed under the ‘Site Validation’ tab of the Model Report, as both a direct comparison of cover estimates as well as the error margin between human and machine annotations for each benthic group. {% hint style="info" %} **Note** that the Project Dashboard and Model Report summaries will ONLY appear once a label set has been fully mapped. {% endhint %} 1. To complete Label Set mapping, on the Label Sets page, click "**Mapping**".

Map labels to the global standard using the Mapping tool.

2. **Filter for Labels** using the searchable filter at the top of the page. 3. Choose a "**Benthic Category**" from the dropdown list for each Label that is the closest taxonomic match to what you are annotating. The Benthic Category list has been adapted from the [World Register of Marine Species](https://www.marinespecies.org/) (WoRMS) with relevant abiotic categories included for completeness. You will need to choose a benthic category that best fits your label *for every single label in your Label Set* to make use of the AI validation and reporting tools, as well as to contribute your data (if you have opted to [share summary data](/get-results/public-dashboard.md)) to regional and global trends analysis. {% hint style="info" %} If you are using one of the AIMS Template Label Sets, the benthic mapping has already been done for you, and will only need to be updated if you add new labels to the Label Set within your Project. {% endhint %}

Mapping your Label Set so that ReefCloud can understand and make use of your data for synthesis is an essential step!

{% hint style="success" %} There are also options to add information about **Growth Form, Substrate** and **Condition**, to add detail to understanding your labels. {% endhint %} --- # Agent Instructions: Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter: ``` GET https://docs.reefcloud.ai/train-and-validate-ai/label-set-editing-and-mapping.md?ask= ``` The question should be specific, self-contained, and written in natural language. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.