A Fusion of Stacking with Dynamic Integration

NF Rooney, WD Patterson

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

    3 Citations (Scopus)

    Abstract

    In this paper we present a novel method that fuses the ensemble meta-techniques of Stack-ing and Dynamic Integration (DI) for regres-sion problems, without adding any major computational overhead. The intention of the technique is to benefit from the varying per-formance of Stacking and DI for different da-ta sets, in order to provide a more robust technique. We detail an empirical analysis of the technique referred to as weighted Meta –Combiner (wMetaComb) and compare its performance to Stacking and the DI technique of Dynamic Weighting with Selection. The empirical analysis consisted of four sets of experiments where each experiment recorded the cross-fold evaluation of each technique for a large number of diverse data sets, where each base model is created using random feature selection and the same base learning al-gorithm. Each experiment differed in terms of the latter base learning algorithm used. We demonstrate that for each evaluation, wMetaComb was able to outperform DI and Stacking for each experiment and as such fuses the two underlying mechanisms suc-cessfully.
    LanguageEnglish
    Title of host publicationUnknown Host Publication
    Pages2844-2849
    Number of pages6
    Publication statusPublished - Jan 2007
    Event20th International Joint Conference on Artificial Intelligence (IJCAI-07) - Hyderabad, India
    Duration: 1 Jan 2007 → …

    Conference

    Conference20th International Joint Conference on Artificial Intelligence (IJCAI-07)
    Period1/01/07 → …

    Fingerprint

    Fusion reactions
    Electric fuses
    Experiments
    Learning algorithms
    Feature extraction

    Cite this

    Rooney, NF., & Patterson, WD. (2007). A Fusion of Stacking with Dynamic Integration. In Unknown Host Publication (pp. 2844-2849)
    Rooney, NF ; Patterson, WD. / A Fusion of Stacking with Dynamic Integration. Unknown Host Publication. 2007. pp. 2844-2849
    @inproceedings{8b4625e89b5b46e180a3dfdc28b4502d,
    title = "A Fusion of Stacking with Dynamic Integration",
    abstract = "In this paper we present a novel method that fuses the ensemble meta-techniques of Stack-ing and Dynamic Integration (DI) for regres-sion problems, without adding any major computational overhead. The intention of the technique is to benefit from the varying per-formance of Stacking and DI for different da-ta sets, in order to provide a more robust technique. We detail an empirical analysis of the technique referred to as weighted Meta –Combiner (wMetaComb) and compare its performance to Stacking and the DI technique of Dynamic Weighting with Selection. The empirical analysis consisted of four sets of experiments where each experiment recorded the cross-fold evaluation of each technique for a large number of diverse data sets, where each base model is created using random feature selection and the same base learning al-gorithm. Each experiment differed in terms of the latter base learning algorithm used. We demonstrate that for each evaluation, wMetaComb was able to outperform DI and Stacking for each experiment and as such fuses the two underlying mechanisms suc-cessfully.",
    author = "NF Rooney and WD Patterson",
    year = "2007",
    month = "1",
    language = "English",
    pages = "2844--2849",
    booktitle = "Unknown Host Publication",

    }

    Rooney, NF & Patterson, WD 2007, A Fusion of Stacking with Dynamic Integration. in Unknown Host Publication. pp. 2844-2849, 20th International Joint Conference on Artificial Intelligence (IJCAI-07), 1/01/07.

    A Fusion of Stacking with Dynamic Integration. / Rooney, NF; Patterson, WD.

    Unknown Host Publication. 2007. p. 2844-2849.

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

    TY - GEN

    T1 - A Fusion of Stacking with Dynamic Integration

    AU - Rooney, NF

    AU - Patterson, WD

    PY - 2007/1

    Y1 - 2007/1

    N2 - In this paper we present a novel method that fuses the ensemble meta-techniques of Stack-ing and Dynamic Integration (DI) for regres-sion problems, without adding any major computational overhead. The intention of the technique is to benefit from the varying per-formance of Stacking and DI for different da-ta sets, in order to provide a more robust technique. We detail an empirical analysis of the technique referred to as weighted Meta –Combiner (wMetaComb) and compare its performance to Stacking and the DI technique of Dynamic Weighting with Selection. The empirical analysis consisted of four sets of experiments where each experiment recorded the cross-fold evaluation of each technique for a large number of diverse data sets, where each base model is created using random feature selection and the same base learning al-gorithm. Each experiment differed in terms of the latter base learning algorithm used. We demonstrate that for each evaluation, wMetaComb was able to outperform DI and Stacking for each experiment and as such fuses the two underlying mechanisms suc-cessfully.

    AB - In this paper we present a novel method that fuses the ensemble meta-techniques of Stack-ing and Dynamic Integration (DI) for regres-sion problems, without adding any major computational overhead. The intention of the technique is to benefit from the varying per-formance of Stacking and DI for different da-ta sets, in order to provide a more robust technique. We detail an empirical analysis of the technique referred to as weighted Meta –Combiner (wMetaComb) and compare its performance to Stacking and the DI technique of Dynamic Weighting with Selection. The empirical analysis consisted of four sets of experiments where each experiment recorded the cross-fold evaluation of each technique for a large number of diverse data sets, where each base model is created using random feature selection and the same base learning al-gorithm. Each experiment differed in terms of the latter base learning algorithm used. We demonstrate that for each evaluation, wMetaComb was able to outperform DI and Stacking for each experiment and as such fuses the two underlying mechanisms suc-cessfully.

    UR - http://www.ijcai.org/papers07/contents.php

    UR - http://www.ijcai.org/papers07/contents.php

    M3 - Conference contribution

    SP - 2844

    EP - 2849

    BT - Unknown Host Publication

    ER -

    Rooney NF, Patterson WD. A Fusion of Stacking with Dynamic Integration. In Unknown Host Publication. 2007. p. 2844-2849