> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.nvidia.com/sdgm/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.nvidia.com/sdgm/_mcp/server.

# Link Prediction

> Metrics reported for link prediction tasks, including NDCG, MAP, MRR, and rank-at-K variants.

> **Prerelease documentation**
>
> This Kumo Studio information is prerelease and is not yet supported. Stay tuned for the Docker and Kubernetes releases.

For link prediction tasks, Kumo reports F1\@K, MAP\@K, Precision\@K, and Recall\@K (for K = 1, 10, and 100), as well as the following metrics:

**Normalized Discounted Cumulative Gain (NDCG)** is calculated by:

![NDCG formula](file:images/Screenshot2025-06-27at10.41.36AM.png "NDCG formula")

where **\_reli \_**is the graded relevance of the result at position ***i***.

**Mean Average Precision (MAP)** is calculated by:

![MAP formula](file:images/Screenshot2025-06-27at10.41.57AM.png "MAP formula")

where MAP is calculated by finding the average precision (AP) for each query, and then averaging over the number of queries.

**Mean Reciprocal Rank (MRR)** is calculated by:

![MRR formula](file:images/Screenshot2025-06-27at10.42.26AM.png "MRR formula")

where **\_ranki \_**refers to the rank position of the *first* relevant item for the *i*-th query.