Irisデータを用いたFQHNNの因果の構造とリアルタイム性の一考察

抄録

It is effective for promoting self-driving car society to grasp driver’s emotions in level 3 and to grasp remote observer ‘s ones in level 4 which are difficult for Convolutional Neural Network (CNN) which is leading the present third generation of AI because CNN has not clear causality. And so our object is to grasp these emotions by using Fuzzy quantification(FQ) theory aided Holographic Neural Network(FQHNN) because both of HNN and FQ theory have causality and HNN is highefficiency. But here we try to grasp the causal characteristics of FQHNN by using IRIS data which is often used for classification. As a result, it becomes clear that only HNN has not enough accurate but FQHNN is good accuracy in both learning and predict. And we get two causal characteristics from FQHNN.