One-class document classification via Neural Networks

Larry Manevitz; Malik Yousef

doi:10.1016/j.neucom.2006.05.013

One-class document classification via Neural Networks

Larry Manevitz, Malik Yousef

Department of Computer Science

Research output: Contribution to journal › Article › peer-review

Abstract

Automated document retrieval and classification is of central importance in many contexts; our main motivating goal is the efficient classification and retrieval of "interests" on the internet when only positive information is available. In this paper, we show how a simple feed-forward neural network can be trained to filter documents under these conditions, and that this method seems to be superior to modified methods (modified to use only positive examples), such as Rocchio, Nearest Neighbor, Naive-Bayes, Distance-based Probability and One-Class SVM algorithms. A novel experimental finding is that retrieval is enhanced substantially in this context by carrying out a certain kind of uniform transformation ("Hadamard") of the information prior to the training of the network.

Original language	English
Pages (from-to)	1466-1481
Number of pages	16
Journal	Neurocomputing
Volume	70
Issue number	7-9
DOIs	https://doi.org/10.1016/j.neucom.2006.05.013
State	Published - Mar 2007

Bibliographical note

Funding Information:
This work was partially supported by HIACS, the Haifa Interdisciplinary Center for Advanced Computer Science. This work forms part of the doctoral thesis of the second author who was supported by a University of Haifa fellowship during his studies, and afterwards hosted by the Neurocomputation Laboratory situated in the Caesarea Rothschild Institute for Interdisciplinary Computer Science. We thank Nathalie Japkowicz for lending us some Matlab software which was used during a revision of this paper. The first author thanks Oxford University for its hospitality during his sabbatical visit.

Funding Information:
This work was partially supported by HIACS, the Haifa University Interdisciplinary Center for Advanced Computer Science, and the Neurocomputation Laboratory located at the Caesarea Rothschild Institute for Interdisciplinary Computer Science.

Keywords

Autoencoder
Automated document retrieval
Bottleneck neural network
Classification
Feed-forward neural networks
Machine learning
One-class classification

ASJC Scopus subject areas

Computer Science Applications
Cognitive Neuroscience
Artificial Intelligence

Access to Document

10.1016/j.neucom.2006.05.013

Cite this

@article{065cd529664940ff9125df60bc364e34,

title = "One-class document classification via Neural Networks",

abstract = "Automated document retrieval and classification is of central importance in many contexts; our main motivating goal is the efficient classification and retrieval of {"}interests{"} on the internet when only positive information is available. In this paper, we show how a simple feed-forward neural network can be trained to filter documents under these conditions, and that this method seems to be superior to modified methods (modified to use only positive examples), such as Rocchio, Nearest Neighbor, Naive-Bayes, Distance-based Probability and One-Class SVM algorithms. A novel experimental finding is that retrieval is enhanced substantially in this context by carrying out a certain kind of uniform transformation ({"}Hadamard{"}) of the information prior to the training of the network.",

keywords = "Autoencoder, Automated document retrieval, Bottleneck neural network, Classification, Feed-forward neural networks, Machine learning, One-class classification",

author = "Larry Manevitz and Malik Yousef",

note = "Funding Information: This work was partially supported by HIACS, the Haifa Interdisciplinary Center for Advanced Computer Science. This work forms part of the doctoral thesis of the second author who was supported by a University of Haifa fellowship during his studies, and afterwards hosted by the Neurocomputation Laboratory situated in the Caesarea Rothschild Institute for Interdisciplinary Computer Science. We thank Nathalie Japkowicz for lending us some Matlab software which was used during a revision of this paper. The first author thanks Oxford University for its hospitality during his sabbatical visit. Funding Information: This work was partially supported by HIACS, the Haifa University Interdisciplinary Center for Advanced Computer Science, and the Neurocomputation Laboratory located at the Caesarea Rothschild Institute for Interdisciplinary Computer Science. ",

year = "2007",

month = mar,

doi = "10.1016/j.neucom.2006.05.013",

language = "English",

volume = "70",

pages = "1466--1481",

journal = "Neurocomputing",

issn = "0925-2312",

publisher = "Elsevier",

number = "7-9",

}

TY - JOUR

T1 - One-class document classification via Neural Networks

AU - Manevitz, Larry

AU - Yousef, Malik

N1 - Funding Information: This work was partially supported by HIACS, the Haifa Interdisciplinary Center for Advanced Computer Science. This work forms part of the doctoral thesis of the second author who was supported by a University of Haifa fellowship during his studies, and afterwards hosted by the Neurocomputation Laboratory situated in the Caesarea Rothschild Institute for Interdisciplinary Computer Science. We thank Nathalie Japkowicz for lending us some Matlab software which was used during a revision of this paper. The first author thanks Oxford University for its hospitality during his sabbatical visit. Funding Information: This work was partially supported by HIACS, the Haifa University Interdisciplinary Center for Advanced Computer Science, and the Neurocomputation Laboratory located at the Caesarea Rothschild Institute for Interdisciplinary Computer Science.

PY - 2007/3

Y1 - 2007/3

N2 - Automated document retrieval and classification is of central importance in many contexts; our main motivating goal is the efficient classification and retrieval of "interests" on the internet when only positive information is available. In this paper, we show how a simple feed-forward neural network can be trained to filter documents under these conditions, and that this method seems to be superior to modified methods (modified to use only positive examples), such as Rocchio, Nearest Neighbor, Naive-Bayes, Distance-based Probability and One-Class SVM algorithms. A novel experimental finding is that retrieval is enhanced substantially in this context by carrying out a certain kind of uniform transformation ("Hadamard") of the information prior to the training of the network.

AB - Automated document retrieval and classification is of central importance in many contexts; our main motivating goal is the efficient classification and retrieval of "interests" on the internet when only positive information is available. In this paper, we show how a simple feed-forward neural network can be trained to filter documents under these conditions, and that this method seems to be superior to modified methods (modified to use only positive examples), such as Rocchio, Nearest Neighbor, Naive-Bayes, Distance-based Probability and One-Class SVM algorithms. A novel experimental finding is that retrieval is enhanced substantially in this context by carrying out a certain kind of uniform transformation ("Hadamard") of the information prior to the training of the network.

KW - Autoencoder

KW - Automated document retrieval

KW - Bottleneck neural network

KW - Classification

KW - Feed-forward neural networks

KW - Machine learning

KW - One-class classification

UR - http://www.scopus.com/inward/record.url?scp=33847410597&partnerID=8YFLogxK

U2 - 10.1016/j.neucom.2006.05.013

DO - 10.1016/j.neucom.2006.05.013

M3 - Article

AN - SCOPUS:33847410597

SN - 0925-2312

VL - 70

SP - 1466

EP - 1481

JO - Neurocomputing

JF - Neurocomputing

IS - 7-9

ER -

One-class document classification via Neural Networks

Abstract

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this