While the common scenario nowadays in social networks is the proliferation of offensive language, the focus of attention in this work is the identification of constructive online comments. In order to automatically identify constructive online comments we implement both traditional and deep learning models based on the use of sparse and dense vector semantics. We evaluate these classifiers on a recently created constructive comments corpus comprised of 12,000 annotated news comments, intended to improve the quality of online discussions. The obtained results show how our model based on learning embeddings (dense vectors) is able to match the performance of complicated architectures like recurrent and convolutional neural networks.


    Access

    Check access

    Check availability in my library

    Order at Subito €


    Export, share and cite



    Title :

    Representation Learning for Constructive Comments Classification


    Contributors:


    Publication date :

    2020-11-01


    Size :

    101745 byte





    Type of media :

    Conference paper


    Type of material :

    Electronic Resource


    Language :

    English



    The regulation of constructive learning processes

    Vermunt, J.D. | Online Contents | 1998


    Constructive Criticism

    Emerald Group Publishing | 1956



    Constructive qualification testing

    Garon, R.J. | Engineering Index Backfile | 1955