Reference : Investigating lexical effects in syntax with regularized regression (Lasso).
Scientific journals : Article
Arts & humanities : Languages & linguistics
Investigating lexical effects in syntax with regularized regression (Lasso).
Van de Velde, Freek [> >]
Pijpops, Dirk mailto [Université de Liège - ULiège > Département de langues modernes : ling., litt. et trad. > Lilith > >]
Research Design and Statistics in Linguistics and Communication Science
[en] Lasso ; regression ; lexical effects ; variation ; language ; statistics
[en] Within usage-based theory, notably in construction grammar though also elsewhere, the role of the lexicon and of lexically-specific patterns in morphosyntax is well recognized. The methodology, however, is not always sufficiently suited to get at the details, as lexical effects are difficult to study under what are currently the standard methods for investigating grammar empirically. In this short article, we propose a method from machine learning: regularized regression (Lasso) with k-fold cross-validation, and compare its performance with a Distinctive Collexeme Analysis.
Liège, Literature, Linguistics - ULiège - Lilith

File(s) associated to this reference

Fulltext file(s):

Open access
VandeVelde Pijpops_2021.pdfAuthor preprint314.07 kBView/Open

Bookmark and Share SFX Query

All documents in ORBi are protected by a user license.