thevar1able/ClickHouse
1
0
Fork 0
mirror of https://github.com/ClickHouse/ClickHouse.git synced 2024-11-22 23:52:03 +00:00
Code Issues Packages Projects Releases Wiki Activity

Fix copyright issues in ANN docs

Browse Source
This commit is contained in:
Robert Schulze 2023-08-14 07:36:27 +00:00
parent 385332a554
commit f71ce2641c
No known key found for this signature in database
GPG Key ID: 26703B55FB13728A
1 changed files with 9 additions and 15 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

24
docs/en/engines/table-engines/mergetree-family/annindexes.md
View File

@ -188,23 +188,17 @@ ENGINE = MergeTree
ORDER BY id;
```
Annoy currently supports `L2Distance` and `cosineDistance` as distance function `Distance`. If no distance function was specified during
index creation, `L2Distance` is used as default. Parameter `NumTrees` is the number of trees which the algorithm creates (default if not
specified: 100). Higher values of `NumTree` mean more accurate search results but slower index creation / query times (approximately
linearly) as well as larger index sizes.
Annoy currently supports two distance functions:
- `L2Distance`, also called Euclidean distance is the length of a line segment between two points in Euclidean space
([Wikipedia](https://en.wikipedia.org/wiki/Euclidean_distance)).
- `cosineDistance`, also called cosine similarity, is the cosine of the angle between two (non-zero) vectors
([Wikipedia](https://en.wikipedia.org/wiki/Cosine_similarity)).
`L2Distance` is also called Euclidean distance, the Euclidean distance between two points in Euclidean space is the length of a line segment between the two points.
For example: If we have point P(p1,p2), Q(q1,q2), their distance will be d(p,q)
![L2Distance](https://en.wikipedia.org/wiki/Euclidean_distance#/media/File:Euclidean_distance_2d.svg)
For normalized data, `L2Distance` is usually a better choice, otherwise `cosineDistance` is recommended to compensate for scale. If no
distance function was specified during index creation, `L2Distance` is used as default.
`cosineDistance` also called cosine similarity is a measure of similarity between two non-zero vectors defined in an inner product space. Cosine similarity is the cosine of the angle between the vectors; that is, it is the dot product of the vectors divided by the product of their lengths.
![cosineDistance](https://www.tyrrell4innovation.ca/wp-content/uploads/2021/06/rsz_jenny_du_miword.png)
The Euclidean distance corresponds to the L2-norm of a difference between vectors. The cosine similarity is proportional to the dot product of two vectors and inversely proportional to the product of their magnitudes.
![compare](https://www.researchgate.net/publication/320914786/figure/fig2/AS:558221849841664@1510101868614/The-difference-between-Euclidean-distance-and-cosine-similarity.png)
In one sentence: cosine similarity care only about the angle between them, but do not care about the "distance" we normally think.
![L2 distance](https://www.baeldung.com/wp-content/uploads/sites/4/2020/06/4-1.png)
![cosineDistance](https://www.baeldung.com/wp-content/uploads/sites/4/2020/06/5.png)
Parameter `NumTrees` is the number of trees which the algorithm creates (default if not specified: 100). Higher values of `NumTree` mean
more accurate search results but slower index creation / query times (approximately linearly) as well as larger index sizes.
:::note
Indexes over columns of type `Array` will generally work faster than indexes on `Tuple` columns. All arrays **must** have same length. Use