Pages

Week 5: Ontologies for sensor data

We read:  Making Sense of Sensor Data Using Ontology: A Discussion for Road Vehicle
Classification

Good paper, honest and specific about what it's trying to say.

Not convinced by the argument that you need an ontology for this.  Inference rules are hardcoded
anyway.  But also, why not?

Must've used OWL2.

More useful if it can be linked with other peoples' ontologies.  Otherwise it's just a database with their own labels.

Could check consistency of data.  A plus but not terribly reliable?

Lots of rubbish papers, or shallow papers, that aren't pushing anything forward.  It's difficult to get real insight on a problem (vs. hacking together papers based on other peoples' ideas), but people often aren't willing to spend money on something with unsure results or time scales.

Everything can be an onotology.

"In theory you can just do whatever you think is best, and call it an ontology." - Someone in the room who wishes to remain anonymous.

No comments

Post a Comment

© Ontologies reading group for first year PhDs in CISA
Design:Maira Gall.