A Binary Neural Decision Table Classifier

Victoria Jane Hodge; Simon O'Keefe; Jim Austin

A Binary Neural Decision Table Classifier

Victoria Jane Hodge, Simon O'Keefe, Jim Austin

Computer Science

Research output: Contribution to conference › Paper › peer-review

Abstract

In this paper, we introduce a neural network-based decision table algorithm. We focus on the implementation details of the decision table algorithm when it is constructed using the neural network. Decision tables are simple supervised classifiers which, Kohavi demonstrated, can outperform state-of-the-art classifiers such as C4.5. We couple this power with the efficiency and flexibility of a binary associative-memory neural network. We demonstrate how the binary associative-memory neural network can form the decision table index to map between attribute values and data records. We also show how two attribute selection algorithms, which may be used to pre-select the attributes for the decision table, can easily be implemented within the binary associative-memory neural framework. The first attribute selector uses mutual information between attributes and classes to select the attributes that classify best. The second attribute selector uses a probabilistic approach to evaluate randomly selected attribute subsets.

Original language	English
Publication status	Published - 29 Aug 2004
Event	Proceedings Brain Inspired Cognitive Systems 2004 (BICS 2004) - Stirling, United Kingdom Duration: 29 Aug 2004 → 1 Sept 2004

Conference

Conference	Proceedings Brain Inspired Cognitive Systems 2004 (BICS 2004)
Country/Territory	United Kingdom
City	Stirling
Period	29/08/04 → 1/09/04

Access to Document

Hodge+OKeefe+Austin_BICS04Submitted manuscript, 318 KB

Cite this

@conference{d675396a065c47fc84f019166007eec0,

title = "A Binary Neural Decision Table Classifier",

abstract = "In this paper, we introduce a neural network-based decision table algorithm. We focus on the implementation details of the decision table algorithm when it is constructed using the neural network. Decision tables are simple supervised classifiers which, Kohavi demonstrated, can outperform state-of-the-art classifiers such as C4.5. We couple this power with the efficiency and flexibility of a binary associative-memory neural network. We demonstrate how the binary associative-memory neural network can form the decision table index to map between attribute values and data records. We also show how two attribute selection algorithms, which may be used to pre-select the attributes for the decision table, can easily be implemented within the binary associative-memory neural framework. The first attribute selector uses mutual information between attributes and classes to select the attributes that classify best. The second attribute selector uses a probabilistic approach to evaluate randomly selected attribute subsets.",

author = "Hodge, {Victoria Jane} and Simon O'Keefe and Jim Austin",

year = "2004",

month = aug,

day = "29",

language = "English",

note = "Proceedings Brain Inspired Cognitive Systems 2004 (BICS 2004) ; Conference date: 29-08-2004 Through 01-09-2004",

}

TY - CONF

T1 - A Binary Neural Decision Table Classifier

AU - Hodge, Victoria Jane

AU - O'Keefe, Simon

AU - Austin, Jim

PY - 2004/8/29

Y1 - 2004/8/29

N2 - In this paper, we introduce a neural network-based decision table algorithm. We focus on the implementation details of the decision table algorithm when it is constructed using the neural network. Decision tables are simple supervised classifiers which, Kohavi demonstrated, can outperform state-of-the-art classifiers such as C4.5. We couple this power with the efficiency and flexibility of a binary associative-memory neural network. We demonstrate how the binary associative-memory neural network can form the decision table index to map between attribute values and data records. We also show how two attribute selection algorithms, which may be used to pre-select the attributes for the decision table, can easily be implemented within the binary associative-memory neural framework. The first attribute selector uses mutual information between attributes and classes to select the attributes that classify best. The second attribute selector uses a probabilistic approach to evaluate randomly selected attribute subsets.

AB - In this paper, we introduce a neural network-based decision table algorithm. We focus on the implementation details of the decision table algorithm when it is constructed using the neural network. Decision tables are simple supervised classifiers which, Kohavi demonstrated, can outperform state-of-the-art classifiers such as C4.5. We couple this power with the efficiency and flexibility of a binary associative-memory neural network. We demonstrate how the binary associative-memory neural network can form the decision table index to map between attribute values and data records. We also show how two attribute selection algorithms, which may be used to pre-select the attributes for the decision table, can easily be implemented within the binary associative-memory neural framework. The first attribute selector uses mutual information between attributes and classes to select the attributes that classify best. The second attribute selector uses a probabilistic approach to evaluate randomly selected attribute subsets.

M3 - Paper

T2 - Proceedings Brain Inspired Cognitive Systems 2004 (BICS 2004)

Y2 - 29 August 2004 through 1 September 2004

ER -