IF-002
Selg-Organizing Neural Grove: SONG

○井上浩孝（呉高専）

Recently, deep learning neural networks have been used for practical applications to improve classification accuracy. However, the training time of the deep learning neural networks increases in proportion to the number of layers. On the other hand, the training time of multiple classifier systems (MCS) based on self-generating neural trees extremely quick. In this paper, we propose a novel pruning method for efficient classification for the neural network ensembles and we call this model a self-organizing neural grove (SONG). Experiments have been conducted to compare the pruned MCS with an unpruned MCS, the MCS based on C4.5, and k-nearest neighbor method. The results show that the SONG can improve its classification accuracy as well as reducing the computation cost not only toy problems, but also practical problems.