Semi-supervised learning of text classification on bacterial protein-protein interaction documents

Guixian Xu, Zhendong Niu, Peter Uetz, Xu Gao, Xuping Qin, Hongfang D Liu

Research output: Chapter in Book/Report/Conference proceedingConference contribution

4 Citations (Scopus)

Abstract

Protein-protein interaction (PPI) network is essential to understand the fundamental processes governing cell biology. The mining and curation of PPI knowledge is critical for analyzing high-throughput genomics and proteomics data. Several PPI knowledge bases have been generated through expensive manual curation but far from comprehensive. It is desired to have a document classification system which can classify documents as PPI-related or not PPI-related and therefore assist the mining and curation of PPI knowledge. In order to build document classification systems, an annotated corpus is needed where each document in the corpus is tagged with a label (either positive or negative). However, it is usually the case that only a small number of positive documents can be obtained manually or from existing PPI knowledge bases with literature evidences. Meanwhile, there are a large number of unlabeled documents where most of them are not PPI-related. Machine learning based on a small number of positives and a large number of unlabeled documents is called learning from positive and unlabelled documents (LPU) which has been studied in the general domain. A popular approach for LPU is a two-step strategy where the first step is to obtain reliable negative documents (RN) and the second step is to refine RN using various methods such as clustering or boosting. In this paper, we tackle the problem of LPU for PPI document classification and compare three two-step procedures based on a public data set, Reuters-21578. One is to obtain a negative data set by building a machine learning classifier which treats each unlabelled document as negatives and then classifies unlabelled documents. The second procedure is to refine the negative data set iteratively and consider those unlabeled documents always classified as negative as reliable negatives. The third procedure is to augment the negative data set iteratively by including unlabeled documents classified as negative in any iteration. Three machine learning algorithms were deployed for each two-step procedure.

Original languageEnglish (US)
Title of host publicationProceedings - 2009 International Joint Conference on Bioinformatics, Systems Biology and Intelligent Computing, IJCBS 2009
Pages263-270
Number of pages8
DOIs
StatePublished - 2009
Externally publishedYes
Event2009 International Joint Conference on Bioinformatics, Systems Biology and Intelligent Computing, IJCBS 2009 - Shanghai, China
Duration: Aug 3 2009Aug 5 2009

Other

Other2009 International Joint Conference on Bioinformatics, Systems Biology and Intelligent Computing, IJCBS 2009
CountryChina
CityShanghai
Period8/3/098/5/09

Fingerprint

Supervised learning
Proteins
Learning systems
Bacterial Proteins
Cytology

Keywords

  • Protein-protein interaction
  • Semi-supervised learning
  • Text classification

ASJC Scopus subject areas

  • Software
  • Biomedical Engineering

Cite this

Xu, G., Niu, Z., Uetz, P., Gao, X., Qin, X., & Liu, H. D. (2009). Semi-supervised learning of text classification on bacterial protein-protein interaction documents. In Proceedings - 2009 International Joint Conference on Bioinformatics, Systems Biology and Intelligent Computing, IJCBS 2009 (pp. 263-270). [5260672] https://doi.org/10.1109/IJCBS.2009.68

Semi-supervised learning of text classification on bacterial protein-protein interaction documents. / Xu, Guixian; Niu, Zhendong; Uetz, Peter; Gao, Xu; Qin, Xuping; Liu, Hongfang D.

Proceedings - 2009 International Joint Conference on Bioinformatics, Systems Biology and Intelligent Computing, IJCBS 2009. 2009. p. 263-270 5260672.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Xu, G, Niu, Z, Uetz, P, Gao, X, Qin, X & Liu, HD 2009, Semi-supervised learning of text classification on bacterial protein-protein interaction documents. in Proceedings - 2009 International Joint Conference on Bioinformatics, Systems Biology and Intelligent Computing, IJCBS 2009., 5260672, pp. 263-270, 2009 International Joint Conference on Bioinformatics, Systems Biology and Intelligent Computing, IJCBS 2009, Shanghai, China, 8/3/09. https://doi.org/10.1109/IJCBS.2009.68
Xu G, Niu Z, Uetz P, Gao X, Qin X, Liu HD. Semi-supervised learning of text classification on bacterial protein-protein interaction documents. In Proceedings - 2009 International Joint Conference on Bioinformatics, Systems Biology and Intelligent Computing, IJCBS 2009. 2009. p. 263-270. 5260672 https://doi.org/10.1109/IJCBS.2009.68
Xu, Guixian ; Niu, Zhendong ; Uetz, Peter ; Gao, Xu ; Qin, Xuping ; Liu, Hongfang D. / Semi-supervised learning of text classification on bacterial protein-protein interaction documents. Proceedings - 2009 International Joint Conference on Bioinformatics, Systems Biology and Intelligent Computing, IJCBS 2009. 2009. pp. 263-270
@inproceedings{8e30842860eb46a8a3db58072d9ec5a0,
title = "Semi-supervised learning of text classification on bacterial protein-protein interaction documents",
abstract = "Protein-protein interaction (PPI) network is essential to understand the fundamental processes governing cell biology. The mining and curation of PPI knowledge is critical for analyzing high-throughput genomics and proteomics data. Several PPI knowledge bases have been generated through expensive manual curation but far from comprehensive. It is desired to have a document classification system which can classify documents as PPI-related or not PPI-related and therefore assist the mining and curation of PPI knowledge. In order to build document classification systems, an annotated corpus is needed where each document in the corpus is tagged with a label (either positive or negative). However, it is usually the case that only a small number of positive documents can be obtained manually or from existing PPI knowledge bases with literature evidences. Meanwhile, there are a large number of unlabeled documents where most of them are not PPI-related. Machine learning based on a small number of positives and a large number of unlabeled documents is called learning from positive and unlabelled documents (LPU) which has been studied in the general domain. A popular approach for LPU is a two-step strategy where the first step is to obtain reliable negative documents (RN) and the second step is to refine RN using various methods such as clustering or boosting. In this paper, we tackle the problem of LPU for PPI document classification and compare three two-step procedures based on a public data set, Reuters-21578. One is to obtain a negative data set by building a machine learning classifier which treats each unlabelled document as negatives and then classifies unlabelled documents. The second procedure is to refine the negative data set iteratively and consider those unlabeled documents always classified as negative as reliable negatives. The third procedure is to augment the negative data set iteratively by including unlabeled documents classified as negative in any iteration. Three machine learning algorithms were deployed for each two-step procedure.",
keywords = "Protein-protein interaction, Semi-supervised learning, Text classification",
author = "Guixian Xu and Zhendong Niu and Peter Uetz and Xu Gao and Xuping Qin and Liu, {Hongfang D}",
year = "2009",
doi = "10.1109/IJCBS.2009.68",
language = "English (US)",
isbn = "9780769537399",
pages = "263--270",
booktitle = "Proceedings - 2009 International Joint Conference on Bioinformatics, Systems Biology and Intelligent Computing, IJCBS 2009",

}

TY - GEN

T1 - Semi-supervised learning of text classification on bacterial protein-protein interaction documents

AU - Xu, Guixian

AU - Niu, Zhendong

AU - Uetz, Peter

AU - Gao, Xu

AU - Qin, Xuping

AU - Liu, Hongfang D

PY - 2009

Y1 - 2009

N2 - Protein-protein interaction (PPI) network is essential to understand the fundamental processes governing cell biology. The mining and curation of PPI knowledge is critical for analyzing high-throughput genomics and proteomics data. Several PPI knowledge bases have been generated through expensive manual curation but far from comprehensive. It is desired to have a document classification system which can classify documents as PPI-related or not PPI-related and therefore assist the mining and curation of PPI knowledge. In order to build document classification systems, an annotated corpus is needed where each document in the corpus is tagged with a label (either positive or negative). However, it is usually the case that only a small number of positive documents can be obtained manually or from existing PPI knowledge bases with literature evidences. Meanwhile, there are a large number of unlabeled documents where most of them are not PPI-related. Machine learning based on a small number of positives and a large number of unlabeled documents is called learning from positive and unlabelled documents (LPU) which has been studied in the general domain. A popular approach for LPU is a two-step strategy where the first step is to obtain reliable negative documents (RN) and the second step is to refine RN using various methods such as clustering or boosting. In this paper, we tackle the problem of LPU for PPI document classification and compare three two-step procedures based on a public data set, Reuters-21578. One is to obtain a negative data set by building a machine learning classifier which treats each unlabelled document as negatives and then classifies unlabelled documents. The second procedure is to refine the negative data set iteratively and consider those unlabeled documents always classified as negative as reliable negatives. The third procedure is to augment the negative data set iteratively by including unlabeled documents classified as negative in any iteration. Three machine learning algorithms were deployed for each two-step procedure.

AB - Protein-protein interaction (PPI) network is essential to understand the fundamental processes governing cell biology. The mining and curation of PPI knowledge is critical for analyzing high-throughput genomics and proteomics data. Several PPI knowledge bases have been generated through expensive manual curation but far from comprehensive. It is desired to have a document classification system which can classify documents as PPI-related or not PPI-related and therefore assist the mining and curation of PPI knowledge. In order to build document classification systems, an annotated corpus is needed where each document in the corpus is tagged with a label (either positive or negative). However, it is usually the case that only a small number of positive documents can be obtained manually or from existing PPI knowledge bases with literature evidences. Meanwhile, there are a large number of unlabeled documents where most of them are not PPI-related. Machine learning based on a small number of positives and a large number of unlabeled documents is called learning from positive and unlabelled documents (LPU) which has been studied in the general domain. A popular approach for LPU is a two-step strategy where the first step is to obtain reliable negative documents (RN) and the second step is to refine RN using various methods such as clustering or boosting. In this paper, we tackle the problem of LPU for PPI document classification and compare three two-step procedures based on a public data set, Reuters-21578. One is to obtain a negative data set by building a machine learning classifier which treats each unlabelled document as negatives and then classifies unlabelled documents. The second procedure is to refine the negative data set iteratively and consider those unlabeled documents always classified as negative as reliable negatives. The third procedure is to augment the negative data set iteratively by including unlabeled documents classified as negative in any iteration. Three machine learning algorithms were deployed for each two-step procedure.

KW - Protein-protein interaction

KW - Semi-supervised learning

KW - Text classification

UR - http://www.scopus.com/inward/record.url?scp=70450167556&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=70450167556&partnerID=8YFLogxK

U2 - 10.1109/IJCBS.2009.68

DO - 10.1109/IJCBS.2009.68

M3 - Conference contribution

SN - 9780769537399

SP - 263

EP - 270

BT - Proceedings - 2009 International Joint Conference on Bioinformatics, Systems Biology and Intelligent Computing, IJCBS 2009

ER -