Hello Shahim, I have a very similar use-case. I am trying to map an endoscopy database > where the main entity is a "finding", which is something that is noticed > during an endoscopy procedure. However, the forms that describe findings > (more than one form per finding) have many questions and answers that could > help classify the finding in different ways. However, if I understood your > use-case correctly, I don't want to infer many types, I want to infer the > single most correct/specific type for each finding based on the values > entered in the various forms. More sophisticated classification can then be > done with an OWL reasoner after the appropriate individuals and types are > retrieved. >
Good to know other folks working with medical databases! But do you intend to pick the most specific type directly on the mapping, or you'll let a reasoner do that? In my case, the form values that I might use are normalized into other > tables. There is a table for all the check-box values, a table for the radio > button values, a table to the drop-down values, etc., indexed by finding_id > and form_id. If my classification depends on more than one of these form > values (from two different tables), what is the best way to do this? I see > that a PropertyBridge can have a join specification but a ClassMap can not. > I think it would be easier to specify a "type" condition by using a > PropertyBridge because the "join" specification could simplify writing the > "where" condition. Either way, the mapping appears to be complicated but > maybe a little simpler with PropertyBridge and "join". Otherwise, the > "where" condition in the ClassMap with have to have a full SELECT/JOIN and > see if the resultset is not empty. > I think you can use "join" on ClassMaps, however you should have a property bridge too, linking the joined classes. The answer lies on what's better for inferring knowledge at the step later. I'm assigning many types to simplify the SWRL rules I'm going to use. For I will write Diagnosis(?d), C50(?d), ECIV(?d) -> AdvancedBreastCancer(?d) instead of Diagnosis(?d), C50(?c50), ECIV(?eciv), diagnosisHasAnswer(?d,?c50), diagnosisHasAnswer(?d,?eciv) -> AdvancedBreastCancer(?d) and needing instances of C50 and ECIV, which would lack any semantics. Also, the first solution is easier to write, understand and (hopefully) more computable. dfcp The main problem for me is to deal with a large dataset, i.e., 200M+ rows. > > Thank, > Shahim > > > I was trying to map this on d2r and found two possible solutions: > > 1) create a ClassMap for each condition, or > > 2) create a PropertyBridge for each condition, using the property > rdf:type. > > I can't tell which one is better, considering query performance and > "semantic correctness" (like, am I using correctly PropertyBridge?). But > I've problems in the past trying to making several ClassMaps to one class. > > Thanks in advance! > > -- > diogo patrão > > > > > ------------------------------------------------------------------------------ > Simplify data backup and recovery for your virtual environment with vRanger. > Installation's a snap, and flexible recovery options mean your data is safe, > secure and there when you need it. Discover what all the cheering's about. > Get your free trial download today. http://p.sf.net/sfu/quest-dev2dev2 > > > _______________________________________________ > d2rq-map-devel mailing > [email protected]https://lists.sourceforge.net/lists/listinfo/d2rq-map-devel > > >
------------------------------------------------------------------------------ Simplify data backup and recovery for your virtual environment with vRanger. Installation's a snap, and flexible recovery options mean your data is safe, secure and there when you need it. Discover what all the cheering's about. Get your free trial download today. http://p.sf.net/sfu/quest-dev2dev2
_______________________________________________ d2rq-map-devel mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/d2rq-map-devel
