Last updated: 11 Mar 2026
govuk-document-clustering-experiment: index
layout: default title: Home
Going and being abroad > Travel abroad (1,110 content items)
- Run 1 is the original approach.
- Runs 2-4 use various approaches to reduce the number of topics generated.
- Run 5 allows topic names to use up to 5 words (instead of 3 words).
- Run 6 builds on run 5 by incorporating some context about topics within which the new sub-topics should be named.
| Run | Topic reduction | Max words | Context | No. of topics† | ||||
|---|---|---|---|---|---|---|---|---|
| 1 | None | 3 | None | 22 | Topics & content items | Topics | Documents | Topic hierarchy |
| 2 | Automatic | 3 | None | 7 | Topics & content items | Topics | Documents | Topic hierarchy |
| 3 | Manual (12) | 3 | None | 11 | Topics & content items | Topics | Documents | Topic hierarchy |
| 4 | Manual (6) | 3 | None | 5 | Topics & content items | Topics | Documents | Topic hierarchy |
| 5 | Manual (12) | 5 | None | 11 | Topics & content items | Topics | Documents | Topic hierarchy |
| 6 | Manual (12) | 5 | Custom | 11 | Topics & content items | Topics | Documents | Topic hierarchy |
† Excluding outliers