遺伝的アルゴリズムおよび Support Vector Machine の構造を利用したニューラルネットワークの提案

カテゴリ: 論文誌(論文単位)

グループ名: 【C】電子・情報・システム部門

発行日: 2020/07/01

タイトル(英語): A Novel Neural Network using the Genetic Algorithm and Structure of the Support Vector Machine

著者名: 小川　恭子（大阪府立大学 工学研究科），森　直樹（大阪府立大学 工学研究科）

著者名(英語): Kyoko Ogawa (Osaka Prefecture University), Naoki Mori (Osaka Prefecture University)

キーワード: SVM，遺伝的アルゴリズム，ニューラルネットワーク，SVM-NN　　support vector machine，genetic algorithms，neural network，SVM-NN

要約(英語): Recently, deep learning has been studied as one of the most effective methods in the machine-learning field, and lots of results have been reported. However, the most effective way to construct neural networks has not yet been determined. Besides, the interpretation of an obtained network by a user is difficult. To solve this problem, we have proposed a neural network with a support vector machine (SVM) called “SVM-NN”. In this proposed method, support vectors in the SVM determine the number of neurons in the neural network and their weights and biases. Then, the hyperplane of the neural network is expected to behave similarly to that of the SVM before training. This method has an advantage in that users can understand the mechanism of the network based on the support vectors. However, there are several problems to apply SVM-NN to real problems. In this study, we proposed the SVM-NN with the genetic algorithm. To confirm the effectiveness of proposed methods, the computer simulations are carried out taking benchmark problems as examples.

本誌: 電気学会論文誌C（電子・情報・システム部門誌） Vol.140 No.7 （2020） 特集：2019年電子・情報・システム部門大会

本誌掲載ページ: 810-819 p

原稿種別: 論文／日本語

電子版へのリンク: https://www.jstage.jst.go.jp/article/ieejeiss/140/7/140_810/_article/-char/ja/