Machine intelligence can guide pet dog health pre-diagnosis for casual owner: A neural network approach

Abstract

In this paper, we developed a health pre-diagnosis system with ART2 neural network for Pet dog health monitoring. This system is for the pet owner who does not have deep knowledge on the pet diseases nor computer technology. The standardized database of symptoms/diseases associations is constructed from textbooks and the user simply gives the most unusual symptom that is found from the dog and then the communication between the user and the system refines and expands the input symptoms through queries. Then an unsupervised ART2 learning system checks the similarity between input and stored diseases with confidence and generates three most probable diseases as output. The system has incremental learning ability and learning by experience ability thus appropriately changes the database over time even without user's database update. In spite of the fact that the system is ad-hoc in nature, the system's performance is verified by veterinarian as adequate and it can stimulate the owner's attention on the dog's abnormality in time such that appropriate professional treatment is given in its early stage. © 2014 SERSC.