An animal disease diagnosis system based on the architecture of binary-inference-core

Abstract

in this paper, we propose a binary-inference-core diagnosis mechanism, which based on the two algorithms: one named Weighted Uncertainty Reason Algorithm Supporting Certainty Factor Speculation and another named Improved Bayesian method supporting machine learning. On the basis of that, its corresponding software system prototype is constructed, and some novel terms and algorithms are initiated systematically. Experimental statistics show that in contrast to the AI diagnosis system based on the traditional mono-inference-core, the binary-inference-core system is able to significantly improve inference accuracy and utilization rate of field knowledge, and its accurate rate is over 92%, while it provides contrast of results from different algorithm, presenting an agreeable macro effect of diagnosis.