A Semi-discriminative Approach for Sub-sentence Level Topic Classification on a Small Dataset

C. Ferner; S. Wegenkittl

doi:10.1007/978-3-030-46147-8_42

A Semi-discriminative Approach for Sub-sentence Level Topic Classification on a Small Dataset

C. Ferner^*, S. Wegenkittl

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

This paper aims at identifying sequences of words related to specific product components in online product reviews. A reliable baseline performance for this topic classification problem is given by a Max Entropy classifier which assumes independence over subsequent topics. However, the reviews exhibit an inherent structure on the document level allowing to frame the task as sequence classification problem. Since more flexible models from the class of Conditional Random Fields were not competitive because of the limited amount of training data available, we propose using a Hidden Markov Model instead and decouple the training of transition and emission probabilities. The discriminating power of the Max Entropy approach is used for the latter. Besides outperforming both standalone methods as well as more generic models such as linear-chain Conditional Random Fields, the combined classifier is able to assign topics on sub-sentence level although labeling in the training data is only available on sentence level. © Springer Nature Switzerland AG 2020.

Original language	English
Title of host publication	Machine Learning and Knowledge Discovery in Databases
Subtitle of host publication	European Conference, ECML PKDD 2019, Würzburg, Germany, September 16–20, 2019, Proceedings, Part II
Publisher	Springer
Number of pages	14
Volume	11907 LNAI
ISBN (Electronic)	978-3-030-46147-8
ISBN (Print)	978-3-030-46146-1
DOIs	https://doi.org/10.1007/978-3-030-46147-8_42
Publication status	Published - Apr 2020
Event	European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2019 - Würzburg, Germany Duration: 16 Sept 2019 → 20 Sept 2019 https://ecmlpkdd2019.org/

Conference

Conference	European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2019
Abbreviated title	ECML PKDD 2019
Country/Territory	Germany
City	Würzburg
Period	16/09/19 → 20/09/19
Internet address	https://ecmlpkdd2019.org/

Keywords

Hidden Markov Model
Small data
Topic classification
Entropy
Hidden Markov models
Information retrieval systems
Machine learning
Base-line performance
Combined classifiers
Conditional random field
Discriminating power
Discriminative approach
Emission probabilities
Online product reviews
Sequence classification
Classification (of information)

Access to Document

10.1007/978-3-030-46147-8_42

https://www.scopus.com/inward/record.uri?eid=2-s2.0-85084813327&doi=10.1007%2f978-3-030-46147-8_42&partnerID=40&md5=f0fe0e8ac63ead5732df09f41562b414

Cite this

@inproceedings{4b32a375b9bc40fdaab315223250e681,

title = "A Semi-discriminative Approach for Sub-sentence Level Topic Classification on a Small Dataset",

abstract = "This paper aims at identifying sequences of words related to specific product components in online product reviews. A reliable baseline performance for this topic classification problem is given by a Max Entropy classifier which assumes independence over subsequent topics. However, the reviews exhibit an inherent structure on the document level allowing to frame the task as sequence classification problem. Since more flexible models from the class of Conditional Random Fields were not competitive because of the limited amount of training data available, we propose using a Hidden Markov Model instead and decouple the training of transition and emission probabilities. The discriminating power of the Max Entropy approach is used for the latter. Besides outperforming both standalone methods as well as more generic models such as linear-chain Conditional Random Fields, the combined classifier is able to assign topics on sub-sentence level although labeling in the training data is only available on sentence level. {\textcopyright} Springer Nature Switzerland AG 2020.",

keywords = "Hidden Markov Model, Small data, Topic classification, Entropy, Hidden Markov models, Information retrieval systems, Machine learning, Base-line performance, Combined classifiers, Conditional random field, Discriminating power, Discriminative approach, Emission probabilities, Online product reviews, Sequence classification, Classification (of information)",

author = "C. Ferner and S. Wegenkittl",

note = "Conference code: 239479 Export Date: 14 December 2023 Correspondence Address: Ferner, C.; Salzburg University of Applied Sciences, Urstein Sued 1, Austria; email: [email protected] References: Baldwin, T., de Marneffe, M.C., Han, B., Kim, Y.B., Ritter, A., Xu, W., Shared tasks of the 2015 workshop on noisy user-generated text: Twitter lexical normalization and named entity recognition (2015) Proceedings of the Workshop on Noisy User-Generated Text, pp. 126-135. , pp; Dai, A.M., Le, Q.V., Semi-supervised sequence learning (2015) Advances in Neural Information Processing Systems, pp. 3079-3087. , pp; (2018) Aggregated Text Corpus of Laptop Expert Reviews with Annotated Topics, , https://github.com/factai/corpus-laptop-topic; Klein, D., Manning, C.D., Conditional structure versus conditional estimation in NLP models (2002) Proceedings of the ACL Conference on Empirical Methods in Natural Language Processing, EMNLP 2002, pp. 9-16. , ACL; Lafferty, J.D., McCallum, A., Pereira, F.C.N., Conditional random fields: Probabilistic models for segmenting and labeling sequence data (2001) Proceedings of the Eighteenth International Conference on Machine Learning, ICML, 2001, pp. 282-289. , pp; Manning, C.D., Raghavan, P., Sch{\"u}tze, H., (2008) Introduction to Information Retrieval, 1. , vol., Cambridge University Press, Cambridge; Manning, C.D., Sch{\"u}tze, H., (1999) Foundations of Statistical Natural Language Processing, , MIT Press, Cambridge; McCallum, A., Efficiently inducing features of conditional random fields (2002) Proceedings of the Nineteenth Conference on Uncertainty in Artificial Intelligence, pp. 403-410. , pp., Morgan Kaufmann Publishers Inc; McCallum, A., Freitag, D., Pereira, F.C.N., Maximum entropy Markov models for information extraction and segmentation (2000) Proceedings of the Seventeenth International Conference on Machine Learning, ICML, 2000, pp. 591-598. , pp; Medlock, B.W., (2008) Investigating Classification for Natural Language Processing Tasks, , University of Cambridge, Computer Laboratory, Technical report; Ng, A.Y., Jordan, M.I., On discriminative vs. Generative classifiers: A comparison of logistic regression and naive Bayes (2002) Advances in Neural Information Processing Systems, pp. 841-848. , pp; Pang, B., Lee, L., Vaithyanathan, S., Thumbs Up? Sentiment classification using machine learning techniques (2002) Proceedings of the ACL-02 Conference on Empirical Methods in Natural Language Processing, 10, pp. 79-86. , Volume, pp.,. Association for Computational Linguistics; Pedregosa, F., Scikit-learn: Machine learning in python (2011) J. Mach. Learn. Res., 12, pp. 2825-2830; Peng, F., McCallum, A., Information extraction from research papers using conditional random fields (2006) Inf. Process. Manage., 42 (4), pp. 963-979; Petrushin, V.A., Hidden Markov models: Fundamentals and applications (2000) Online Symposium for Electronics Engineer; Pinto, D., McCallum, A., Wei, X., Croft, W.B., Table extraction using conditional random fields (2003) Proceedings of the 26Th Annual International ACM SIGIR Conference on Research and Development in Informaion Retrieval, pp. 235-242. , pp., ACM; Pontiki, M., SemEval-2016 task 5: Aspect based sentiment analysis (2016) Proceedings of the 10Th International Workshop on Semantic Evaluation (Semeval-2016), pp. 19-30. , pp; Pontiki, M., Galanis, D., Pavlopoulos, J., Papageorgiou, H., Androutsopoulos, I., Manandhar, S., SemEval-2014 task 4: Aspect based sentiment analysis (2014) Proceedings of the 8Th International Workshop on Semantic Evaluation (Semeval-2014), pp. 27-35. , pp; Schwartz, A.S., (2007) Posterior Decoding Methods for Optimization and Accuracy Control of Multiple Alignments, , Ph.D. thesis, EECS Department, University of California, Berkeley; Sha, F., Pereira, F., Shallow parsing with conditional random fields (2003) Proceedings of the 2003 Conference of the North American Chapter of the Association for Computational Linguistics on Human Language Technology, pp. 134-141. , pp., ACL; Strauss, B., Toma, B., Ritter, A., de Marneffe, M.C., Xu, W., Results of the WNUT16 named entity recognition shared task (2016) Proceedings of the 2Nd Workshop on Noisy User-Generated Text (WNUT), pp. 138-144. , pp; Sutton, C., McCallum, A., An introduction to conditional random fields (2012) Found. Trends {\textregistered} Mach. Learn., 4 (4), pp. 267-373; European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2019, ECML PKDD 2019 ; Conference date: 16-09-2019 Through 20-09-2019",

year = "2020",

month = apr,

doi = "10.1007/978-3-030-46147-8\_42",

language = "English",

isbn = "978-3-030-46146-1",

volume = "11907 LNAI",

booktitle = "Machine Learning and Knowledge Discovery in Databases",

publisher = "Springer",

address = "Germany",

url = "https://ecmlpkdd2019.org/",

}

Ferner, C & Wegenkittl, S 2020, A Semi-discriminative Approach for Sub-sentence Level Topic Classification on a Small Dataset. in Machine Learning and Knowledge Discovery in Databases : European Conference, ECML PKDD 2019, Würzburg, Germany, September 16–20, 2019, Proceedings, Part II. vol. 11907 LNAI, Springer, European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2019, Würzburg, Germany, 16/09/19. https://doi.org/10.1007/978-3-030-46147-8_42

A Semi-discriminative Approach for Sub-sentence Level Topic Classification on a Small Dataset. / Ferner, C.; Wegenkittl, S.
Machine Learning and Knowledge Discovery in Databases : European Conference, ECML PKDD 2019, Würzburg, Germany, September 16–20, 2019, Proceedings, Part II. Vol. 11907 LNAI Springer, 2020.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - A Semi-discriminative Approach for Sub-sentence Level Topic Classification on a Small Dataset

AU - Ferner, C.

AU - Wegenkittl, S.

N1 - Conference code: 239479 Export Date: 14 December 2023 Correspondence Address: Ferner, C.; Salzburg University of Applied Sciences, Urstein Sued 1, Austria; email: [email protected] References: Baldwin, T., de Marneffe, M.C., Han, B., Kim, Y.B., Ritter, A., Xu, W., Shared tasks of the 2015 workshop on noisy user-generated text: Twitter lexical normalization and named entity recognition (2015) Proceedings of the Workshop on Noisy User-Generated Text, pp. 126-135. , pp; Dai, A.M., Le, Q.V., Semi-supervised sequence learning (2015) Advances in Neural Information Processing Systems, pp. 3079-3087. , pp; (2018) Aggregated Text Corpus of Laptop Expert Reviews with Annotated Topics, , https://github.com/factai/corpus-laptop-topic; Klein, D., Manning, C.D., Conditional structure versus conditional estimation in NLP models (2002) Proceedings of the ACL Conference on Empirical Methods in Natural Language Processing, EMNLP 2002, pp. 9-16. , ACL; Lafferty, J.D., McCallum, A., Pereira, F.C.N., Conditional random fields: Probabilistic models for segmenting and labeling sequence data (2001) Proceedings of the Eighteenth International Conference on Machine Learning, ICML, 2001, pp. 282-289. , pp; Manning, C.D., Raghavan, P., Schütze, H., (2008) Introduction to Information Retrieval, 1. , vol., Cambridge University Press, Cambridge; Manning, C.D., Schütze, H., (1999) Foundations of Statistical Natural Language Processing, , MIT Press, Cambridge; McCallum, A., Efficiently inducing features of conditional random fields (2002) Proceedings of the Nineteenth Conference on Uncertainty in Artificial Intelligence, pp. 403-410. , pp., Morgan Kaufmann Publishers Inc; McCallum, A., Freitag, D., Pereira, F.C.N., Maximum entropy Markov models for information extraction and segmentation (2000) Proceedings of the Seventeenth International Conference on Machine Learning, ICML, 2000, pp. 591-598. , pp; Medlock, B.W., (2008) Investigating Classification for Natural Language Processing Tasks, , University of Cambridge, Computer Laboratory, Technical report; Ng, A.Y., Jordan, M.I., On discriminative vs. Generative classifiers: A comparison of logistic regression and naive Bayes (2002) Advances in Neural Information Processing Systems, pp. 841-848. , pp; Pang, B., Lee, L., Vaithyanathan, S., Thumbs Up? Sentiment classification using machine learning techniques (2002) Proceedings of the ACL-02 Conference on Empirical Methods in Natural Language Processing, 10, pp. 79-86. , Volume, pp.,. Association for Computational Linguistics; Pedregosa, F., Scikit-learn: Machine learning in python (2011) J. Mach. Learn. Res., 12, pp. 2825-2830; Peng, F., McCallum, A., Information extraction from research papers using conditional random fields (2006) Inf. Process. Manage., 42 (4), pp. 963-979; Petrushin, V.A., Hidden Markov models: Fundamentals and applications (2000) Online Symposium for Electronics Engineer; Pinto, D., McCallum, A., Wei, X., Croft, W.B., Table extraction using conditional random fields (2003) Proceedings of the 26Th Annual International ACM SIGIR Conference on Research and Development in Informaion Retrieval, pp. 235-242. , pp., ACM; Pontiki, M., SemEval-2016 task 5: Aspect based sentiment analysis (2016) Proceedings of the 10Th International Workshop on Semantic Evaluation (Semeval-2016), pp. 19-30. , pp; Pontiki, M., Galanis, D., Pavlopoulos, J., Papageorgiou, H., Androutsopoulos, I., Manandhar, S., SemEval-2014 task 4: Aspect based sentiment analysis (2014) Proceedings of the 8Th International Workshop on Semantic Evaluation (Semeval-2014), pp. 27-35. , pp; Schwartz, A.S., (2007) Posterior Decoding Methods for Optimization and Accuracy Control of Multiple Alignments, , Ph.D. thesis, EECS Department, University of California, Berkeley; Sha, F., Pereira, F., Shallow parsing with conditional random fields (2003) Proceedings of the 2003 Conference of the North American Chapter of the Association for Computational Linguistics on Human Language Technology, pp. 134-141. , pp., ACL; Strauss, B., Toma, B., Ritter, A., de Marneffe, M.C., Xu, W., Results of the WNUT16 named entity recognition shared task (2016) Proceedings of the 2Nd Workshop on Noisy User-Generated Text (WNUT), pp. 138-144. , pp; Sutton, C., McCallum, A., An introduction to conditional random fields (2012) Found. Trends ® Mach. Learn., 4 (4), pp. 267-373

PY - 2020/4

Y1 - 2020/4

N2 - This paper aims at identifying sequences of words related to specific product components in online product reviews. A reliable baseline performance for this topic classification problem is given by a Max Entropy classifier which assumes independence over subsequent topics. However, the reviews exhibit an inherent structure on the document level allowing to frame the task as sequence classification problem. Since more flexible models from the class of Conditional Random Fields were not competitive because of the limited amount of training data available, we propose using a Hidden Markov Model instead and decouple the training of transition and emission probabilities. The discriminating power of the Max Entropy approach is used for the latter. Besides outperforming both standalone methods as well as more generic models such as linear-chain Conditional Random Fields, the combined classifier is able to assign topics on sub-sentence level although labeling in the training data is only available on sentence level. © Springer Nature Switzerland AG 2020.

AB - This paper aims at identifying sequences of words related to specific product components in online product reviews. A reliable baseline performance for this topic classification problem is given by a Max Entropy classifier which assumes independence over subsequent topics. However, the reviews exhibit an inherent structure on the document level allowing to frame the task as sequence classification problem. Since more flexible models from the class of Conditional Random Fields were not competitive because of the limited amount of training data available, we propose using a Hidden Markov Model instead and decouple the training of transition and emission probabilities. The discriminating power of the Max Entropy approach is used for the latter. Besides outperforming both standalone methods as well as more generic models such as linear-chain Conditional Random Fields, the combined classifier is able to assign topics on sub-sentence level although labeling in the training data is only available on sentence level. © Springer Nature Switzerland AG 2020.

KW - Hidden Markov Model

KW - Small data

KW - Topic classification

KW - Entropy

KW - Hidden Markov models

KW - Information retrieval systems

KW - Machine learning

KW - Base-line performance

KW - Combined classifiers

KW - Conditional random field

KW - Discriminating power

KW - Discriminative approach

KW - Emission probabilities

KW - Online product reviews

KW - Sequence classification

KW - Classification (of information)

U2 - 10.1007/978-3-030-46147-8_42

DO - 10.1007/978-3-030-46147-8_42

M3 - Conference contribution

SN - 978-3-030-46146-1

VL - 11907 LNAI

BT - Machine Learning and Knowledge Discovery in Databases

PB - Springer

T2 - European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, ECML PKDD 2019

Y2 - 16 September 2019 through 20 September 2019

ER -

A Semi-discriminative Approach for Sub-sentence Level Topic Classification on a Small Dataset

Abstract

Conference

Keywords

Access to Document

Fingerprint

Cite this