高時間分解ACナノポア測定と機械学習による網羅的な細菌の同定

抄録

In this study, we evaluated the performance of a comprehensive bacterial identification method that combines AC nanopore methods and machine learning. The nanopore device for bacteria was fabricated, and the classification performance was evaluated for 11 kinds of bacteria. The classification accuracy using the convolutional neural network (CNN) was 96.6%, and that using the random forest method was 41.5%. These results indicate that the combination of the AC nanopore method and CNN is a promising sensing method that can classify bacteria with high accuracy.