Generation of attributes for learning algorithms

Yuh-Jyh Hu; Dennis Kibler

Generation of attributes for learning algorithms

Yuh-Jyh Hu^*, Dennis Kibler

^*此作品的通信作者

數據科學與工程研究所

研究成果: Paper › 同行評審

31 引文斯高帕斯（Scopus）

摘要

Inductive algorithms rely strongly on their representational biases. Constructive induction can mitigate representational inadequacies. This paper introduces the notion of a relative gain measure and describes a new constructive induction algorithm (GALA) which is independent of the learning algorithm. Unlike most previous research on constructive induction, our methods are designed as preprocessing step before standard machine learning algorithms are applied. We present the results which demonstrate the effectiveness of GALA on artificial and real domains for several learners: C4.5, CN2, perceptron and backpropagation.

原文	English
頁面	806-811
頁數	6
出版狀態	Published - 1 12月 1996
事件	Proceedings of the 1996 13th National Conference on Artificial Intelligence, AAAI 96. Part 1 (of 2) - Portland, OR, USA 持續時間: 4 8月 1996 → 8 8月 1996

Conference

Conference	Proceedings of the 1996 13th National Conference on Artificial Intelligence, AAAI 96. Part 1 (of 2)
城市	Portland, OR, USA
期間	4/08/96 → 8/08/96

其它文件與鏈接

Link to publication in Scopus

引用此

@conference{e5475f1b452e4b14b794af60bf63561d,

title = "Generation of attributes for learning algorithms",

abstract = "Inductive algorithms rely strongly on their representational biases. Constructive induction can mitigate representational inadequacies. This paper introduces the notion of a relative gain measure and describes a new constructive induction algorithm (GALA) which is independent of the learning algorithm. Unlike most previous research on constructive induction, our methods are designed as preprocessing step before standard machine learning algorithms are applied. We present the results which demonstrate the effectiveness of GALA on artificial and real domains for several learners: C4.5, CN2, perceptron and backpropagation.",

author = "Yuh-Jyh Hu and Dennis Kibler",

year = "1996",

month = dec,

day = "1",

language = "English",

pages = "806--811",

note = "Proceedings of the 1996 13th National Conference on Artificial Intelligence, AAAI 96. Part 1 (of 2) ; Conference date: 04-08-1996 Through 08-08-1996",

}

TY - CONF

T1 - Generation of attributes for learning algorithms

AU - Hu, Yuh-Jyh

AU - Kibler, Dennis

PY - 1996/12/1

Y1 - 1996/12/1

N2 - Inductive algorithms rely strongly on their representational biases. Constructive induction can mitigate representational inadequacies. This paper introduces the notion of a relative gain measure and describes a new constructive induction algorithm (GALA) which is independent of the learning algorithm. Unlike most previous research on constructive induction, our methods are designed as preprocessing step before standard machine learning algorithms are applied. We present the results which demonstrate the effectiveness of GALA on artificial and real domains for several learners: C4.5, CN2, perceptron and backpropagation.

AB - Inductive algorithms rely strongly on their representational biases. Constructive induction can mitigate representational inadequacies. This paper introduces the notion of a relative gain measure and describes a new constructive induction algorithm (GALA) which is independent of the learning algorithm. Unlike most previous research on constructive induction, our methods are designed as preprocessing step before standard machine learning algorithms are applied. We present the results which demonstrate the effectiveness of GALA on artificial and real domains for several learners: C4.5, CN2, perceptron and backpropagation.

UR - http://www.scopus.com/inward/record.url?scp=0030361831&partnerID=8YFLogxK

M3 - Paper

AN - SCOPUS:0030361831

SP - 806

EP - 811

T2 - Proceedings of the 1996 13th National Conference on Artificial Intelligence, AAAI 96. Part 1 (of 2)

Y2 - 4 August 1996 through 8 August 1996

ER -

Generation of attributes for learning algorithms

摘要

Conference

其它文件與鏈接

指紋

引用此