star star

Dog Breed Ontology



< Back to course page

Dog Breed Ontology

Design

Research

Resources

Contact

Ontologies

The Dog Breed Ontology was developed with the help of several published ontologies. Below are the ontologies that were used a significant amount in the design and implementation of the system. All ontologies used are listed on the Concept Map & Ontology page of our website.

Several other ontologies exist that aim to describe the characteristics of animals that were not used in this project. They are listed below.

Other Breed Recommendation Systems

Many other systems exist for recommending dog breeds. Our project was inspired by these quizes/questionnaires and aimed to combine the best aspects of each systems with semantic technology.

There were some questions that other recommenders asked about that our system does not account for. Our system has no way to model an experienced vs inexperienced dog owner, and no way to model pets that are not cats and dogs. Other recommenders asked about level of experience in dog ownership and usually allowed for a catch-all category of pets such as ‘other small mammal’ or ‘other small animal.’ It was unclear what data was being used to categorize breeds as ‘good for inexperienced owners’ or ‘good for small animal’ so we did not include those categories.

We ran an experiment to compare the results of our recommender with the other recommenders. We used the same situation for each: Adopter already has a dog and multiple small children, one of whom is mildly allergic to dogs. They live in an apartment and don’t have much time to clean, but is willing to give the dog basic obedience training. We encoded this as the following characteristics that we wanted in a dog:

When the recommenders asked about preferences that are not included in the above description, no preference or the most neutral option was selected. The top results from each recommender were:

References