A deep neural network based hierarchical multi-label classifier for protein function prediction

Xin Yuan, Weite Li, Kui Lin, Jinglu Hu

研究成果: Conference contribution

抄録

Hierarchical classification approaches have been shown to be effective for protein function prediction problem. Traditionally, a set of simple classifiers are used for each hierarchical level separately. In this paper, we introduce a deep neural network (DNN) model with multiple heads and multiple ends to realize the whole set of classifiers. The DNN model consists of three parts: the body part, the multi-end part and the multi-head part. The body part is a deep multilayer perceptron (MLP) shared by different classifiers for feature mapping. The multi-end part performs feature fusion transforming the input vectors of different classifiers to feature vectors with the same length so as to share the feature mapping part. The multi-head part is a set of linear multi-label classifiers. By sharing a deep MLP with multiple classifiers, we are able to construct more powerful classifiers for each level with limited training samples, and expecting to have better classification performance. Experiment results on benchmark datasets show that the proposed method significantly outperforms these traditional methods.

本文言語English
ホスト出版物のタイトルCITS 2019 - Proceeding of the 2019 International Conference on Computer, Information and Telecommunication Systems
編集者Mohammad S. Obaidat, Zhenqiang Mi, Kuei-Fang Hsiao, Petros Nicopolitidis, Daniel Cascado-Caballero
出版社Institute of Electrical and Electronics Engineers Inc.
ISBN(電子版)9781538640883
DOI
出版ステータスPublished - 2019 8
イベント2019 International Conference on Computer, Information and Telecommunication Systems, CITS 2019 - Beijing, China
継続期間: 2019 8 282019 8 31

出版物シリーズ

名前CITS 2019 - Proceeding of the 2019 International Conference on Computer, Information and Telecommunication Systems

Conference

Conference2019 International Conference on Computer, Information and Telecommunication Systems, CITS 2019
国/地域China
CityBeijing
Period19/8/2819/8/31

ASJC Scopus subject areas

  • コンピュータ ビジョンおよびパターン認識
  • 情報システム
  • 情報システムおよび情報管理
  • 安全性、リスク、信頼性、品質管理
  • コンピュータ ネットワークおよび通信
  • ハードウェアとアーキテクチャ

フィンガープリント

「A deep neural network based hierarchical multi-label classifier for protein function prediction」の研究トピックを掘り下げます。これらがまとまってユニークなフィンガープリントを構成します。

引用スタイル