[Archivesspace_Users_Group] Relevance ranking prioritising info in agent record's biog history

[Natalie Adams]

Dear Andrew,

Many thanks for your reply and for the helpful pointers!

Best wishes,

Natalie

Natalie Adams
Systems Archivist
Cambridge University Library


Information about using Cambridge University Libraries is available online here: https://www.lib.cam.ac.uk/using-library


________________________________
From: archivesspace_users_group-bounces at lyralists.lyrasis.org <archivesspace_users_group-bounces at lyralists.lyrasis.org> on behalf of Andrew Morrison <andrew.morrison at bodleian.ox.ac.uk>
Sent: 11 February 2021 11:32
To: archivesspace_users_group at lyralists.lyrasis.org <archivesspace_users_group at lyralists.lyrasis.org>
Subject: Re: [Archivesspace_Users_Group] Relevance ranking prioritising info in agent record's biog history


The default "Keyword" search option causes ArchivesSpace to query a single Solr index field called "fullrecord". That is a big, unstructured free-text field containing everything the indexer deems relevant to each record (and, as in this case, pulled in from other things linked to it, such as agents.) As such, no priority is given to any one note or text field.


Why your example search for "suffrage" is returning that particular record in third place is difficult to say. Have you set up any boost (in config.rb or a plug-in) to increase the relevance score of resources relative to other record types? That would explain why it is in the top six, and why they're all resources. Then other factors would come into play to determine the order among resources, such as term frequency and relative size of each record (the fewer distinct words, the more relevant Solr regards a match within them to be.)


Andrew.



On 11/02/2021 10:39, Natalie Adams wrote:
Dear all,

Now that we have completed the process of migrating legacy catalogue records to ArchivesSpace we have a large number of collections of personal papers available on our ArchivesSpace PUI (https://archivesearch.lib.cam.ac.uk/). Some of them are of very prominent people who have correspondingly long biographical history notes in their agent record. Not all the information in a biographical history is necessarily reflected in the contents of the archival collection- e.g. for a collection of papers of a Nobel prize winning scientist, the biog history would include information about the scientific breakthrough that led to the award of the Nobel even though the papers we have might relate solely to the scientist's later work as Head of one of the Cambridge colleges.

 We have noticed that relevance ranking of search results prioritises information in the biographical history note of an agent record, and will return hits tagged with an agent record (whether or not the search term is found in the resource or archival description data). In a search for suffrage (https://archivesearch.lib.cam.ac.uk/search?utf8=%E2%9C%93&op%5B%5D=&q%5B%5D=suffrage&limit=&field%5B%5D=&from_year%5B%5D=&to_year%5B%5D=&commit=Search) on our system, the third hit in order of relevance (https://archivesearch.lib.cam.ac.uk/repositories/2/resources/7120) appears to have no relevance to the search unless you visit the agent record (https://archivesearch.lib.cam.ac.uk/agents/people/8044) where the term suffrage is included.

We would be very interested to know whether other repositories have experienced the same issue and have considered or been able to make any adjustments to their systems.

Best wishes,

Natalie
ArchiveSearch | ArchiveSearch<https://archivesearch.lib.cam.ac.uk/>
What are archives? Archives are records produced by individuals, families, businesses or organisations during their existence. Archives come in a huge variety of formats- from parchment documents to digital files.
archivesearch.lib.cam.ac.uk


Natalie Adams
Systems Archivist
Cambridge University Library


Information about using Cambridge University Libraries is available online here: https://www.lib.cam.ac.uk/using-library





_______________________________________________
Archivesspace_Users_Group mailing list
Archivesspace_Users_Group at lyralists.lyrasis.org<mailto:Archivesspace_Users_Group at lyralists.lyrasis.org>
http://lyralists.lyrasis.org/mailman/listinfo/archivesspace_users_group

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lyralists.lyrasis.org/pipermail/archivesspace_users_group/attachments/20210211/78f79842/attachment.html>