chin-rcip / collections-model

Linked Open Data Development at the Canadian Heritage Information Network - Développement en données ouvertes et liées au Réseau canadien d'information sur le patrimoine
Creative Commons Zero v1.0 Universal
12 stars 1 forks source link

Modeling uncertainty #47

Open stephenhart8 opened 4 years ago

stephenhart8 commented 4 years ago

In the present 2.1 Target Model, there is no pattern to model uncertainty in the data. By uncertainty, I mean data that is existing, but unsure about its certainty. It is different from the absence of data (see issue #22). This uncertainty about one information could mean that one or multiple uncertain options, that could be inclusive or exclusive.

Let's take for example the address of John. In my opinion, there are 3 different possibilities of uncertainty:

It would be also interesting to add some coefficients to the certainty. For example, the certainty of John living in Montreal is 80% and living in Ottawa 20%.

As I've said in issue #22, I've come across some interesting researches by the Antike Fundmünzel in Europa about managing uncertain data, published at the CAA:

They propose multiple solutions, including:

Also in issue #22, @VladimirAlexiev added the following links:


I have not yet thought of a possible mapping pattern for this topic, but I leave this issue here for further discussions.

VladimirAlexiev commented 4 years ago

My links above are specifically about uncertainty of authorship. Here's how we modeled similar things for the British Museum in ResearchSpace in 2013-2014: https://confluence.ontotext.com/display/ResearchSpace/BM+Association+Mapping+v2#BMAssociationMappingv2-ProducedByCloselyRelatedGroup