Sorting assets by "Relevance"
I’ve got a question regarding the search in the DAM. What is the logic when sorting results by “Relevance”? How does the system decide which asset is more relevant?
A user searches by family, beach and summer and gets 1305 results.
The first result contains the three tags.
The second (highlighted) only summer and beach, no family.
The third only family and summer, no beach.
Further below, there are many assets that have all three tags. Why are the second and third result more “relevant” than others that have all the tags I am searching?
The most relevant for a user would be:
1) assets that contain all three words
2) assets that contain the first and second
3) assets that contain the first and third
4) assets that contain the second and third
5) assets that contain only the first
6) assets that contain only the second
7) assets that contain only the third
-
Hi Rafael,
We are currently using Elasticsearch as our search engine of choice. The algorithm uses different criteria to score an asset’s relevance. These include term frequency, field length norm and other various factors. For example, if an asset includes all three relevant tags but also has other tags on it that were not included in the search, the asset will be of less relevance than an asset that includes only the three or two tags that were searched for.
I hope this better explains the search results. If you have any further questions, please feel free to reach out to us at support team at support@cloudinary.com.
Best regards,
Millie Axelrod
0 -
Hi Millie,
Thanks for the explanation. Although it is not the ideal behaviour for us in terms of searching for the most relevant assets, I must admit it makes sense and it is just as valid :) In any case, it is good to know how it works in order to understand the reason behind the results.
If there is ever a change in the criteria for sorting by relevance, please, consider my suggestion above.
0
Post is closed for comments.
Comments
2 comments