Rule meta-learning for trigram-based sequence processing

S.V.M. Canisius, A. van den Bosch, W. Daelemans

    Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

    Abstract

    Predicting overlapping trigrams of class labels is a recently-proposed method to improve performance on sequence labelling tasks. In this method, sequence elements are effectively classified three times, therefore some procedure is needed to post-process those overlapping classifications into one output sequence. In this paper, we present a rule-based procedure learned automatically from training data. In combination with a memory-based leaner predicting class trigrams, the performance of this meta-learned overlapping trigram post-processor matches that of a handcrafted post-processing rule used in the original study on class trigrams. Moreover, on two domain-specific entity chunking tasks, the class trigram method with automatically learned post-processing rules compares favourably with recent probabilistic sequence labelling techniques, such as maximum-entropy markov models and conditional random fields.
    Original languageEnglish
    Title of host publicationProceedings of the Fourth Learning Language in Logic Workshop
    EditorsJ. Cussens, C. Nedellec
    Place of PublicationBonn
    Publisher[s.n.]
    Pages3-10
    Number of pages8
    Publication statusPublished - 2005

    Fingerprint

    Labeling
    Processing
    Labels
    Entropy
    Data storage equipment

    Cite this

    Canisius, S. V. M., van den Bosch, A., & Daelemans, W. (2005). Rule meta-learning for trigram-based sequence processing. In J. Cussens, & C. Nedellec (Eds.), Proceedings of the Fourth Learning Language in Logic Workshop (pp. 3-10). Bonn: [s.n.].
    Canisius, S.V.M. ; van den Bosch, A. ; Daelemans, W. / Rule meta-learning for trigram-based sequence processing. Proceedings of the Fourth Learning Language in Logic Workshop. editor / J. Cussens ; C. Nedellec. Bonn : [s.n.], 2005. pp. 3-10
    @inproceedings{9d3f6d852b34440da1107ec8818f4a5f,
    title = "Rule meta-learning for trigram-based sequence processing",
    abstract = "Predicting overlapping trigrams of class labels is a recently-proposed method to improve performance on sequence labelling tasks. In this method, sequence elements are effectively classified three times, therefore some procedure is needed to post-process those overlapping classifications into one output sequence. In this paper, we present a rule-based procedure learned automatically from training data. In combination with a memory-based leaner predicting class trigrams, the performance of this meta-learned overlapping trigram post-processor matches that of a handcrafted post-processing rule used in the original study on class trigrams. Moreover, on two domain-specific entity chunking tasks, the class trigram method with automatically learned post-processing rules compares favourably with recent probabilistic sequence labelling techniques, such as maximum-entropy markov models and conditional random fields.",
    author = "S.V.M. Canisius and {van den Bosch}, A. and W. Daelemans",
    note = "Pagination: 8",
    year = "2005",
    language = "English",
    pages = "3--10",
    editor = "J. Cussens and C. Nedellec",
    booktitle = "Proceedings of the Fourth Learning Language in Logic Workshop",
    publisher = "[s.n.]",

    }

    Canisius, SVM, van den Bosch, A & Daelemans, W 2005, Rule meta-learning for trigram-based sequence processing. in J Cussens & C Nedellec (eds), Proceedings of the Fourth Learning Language in Logic Workshop. [s.n.], Bonn, pp. 3-10.

    Rule meta-learning for trigram-based sequence processing. / Canisius, S.V.M.; van den Bosch, A.; Daelemans, W.

    Proceedings of the Fourth Learning Language in Logic Workshop. ed. / J. Cussens; C. Nedellec. Bonn : [s.n.], 2005. p. 3-10.

    Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

    TY - GEN

    T1 - Rule meta-learning for trigram-based sequence processing

    AU - Canisius, S.V.M.

    AU - van den Bosch, A.

    AU - Daelemans, W.

    N1 - Pagination: 8

    PY - 2005

    Y1 - 2005

    N2 - Predicting overlapping trigrams of class labels is a recently-proposed method to improve performance on sequence labelling tasks. In this method, sequence elements are effectively classified three times, therefore some procedure is needed to post-process those overlapping classifications into one output sequence. In this paper, we present a rule-based procedure learned automatically from training data. In combination with a memory-based leaner predicting class trigrams, the performance of this meta-learned overlapping trigram post-processor matches that of a handcrafted post-processing rule used in the original study on class trigrams. Moreover, on two domain-specific entity chunking tasks, the class trigram method with automatically learned post-processing rules compares favourably with recent probabilistic sequence labelling techniques, such as maximum-entropy markov models and conditional random fields.

    AB - Predicting overlapping trigrams of class labels is a recently-proposed method to improve performance on sequence labelling tasks. In this method, sequence elements are effectively classified three times, therefore some procedure is needed to post-process those overlapping classifications into one output sequence. In this paper, we present a rule-based procedure learned automatically from training data. In combination with a memory-based leaner predicting class trigrams, the performance of this meta-learned overlapping trigram post-processor matches that of a handcrafted post-processing rule used in the original study on class trigrams. Moreover, on two domain-specific entity chunking tasks, the class trigram method with automatically learned post-processing rules compares favourably with recent probabilistic sequence labelling techniques, such as maximum-entropy markov models and conditional random fields.

    M3 - Conference contribution

    SP - 3

    EP - 10

    BT - Proceedings of the Fourth Learning Language in Logic Workshop

    A2 - Cussens, J.

    A2 - Nedellec, C.

    PB - [s.n.]

    CY - Bonn

    ER -

    Canisius SVM, van den Bosch A, Daelemans W. Rule meta-learning for trigram-based sequence processing. In Cussens J, Nedellec C, editors, Proceedings of the Fourth Learning Language in Logic Workshop. Bonn: [s.n.]. 2005. p. 3-10