The impact of the frequency-by-regularity interaction on predictive processing
收藏资源简介：
Title: The impact of the frequency-by-regularity interaction on predictive processing Author: Claudia MarziAffiliation: Institute for Computational Linguistics "A. Zampolli", National Research Council of Italy email: claudia.marzi@cnr.it Journal: Word Structure (forthcoming) Description:These datasets and dataframes contain model (TSOM, Temporal self-organisng map) inputs and outputs for simulations of morphological learning and predictive processing in Italian verb forms. The data are organised by paradigm type (regular vs. irregular) and training regime (corpus-based, uniform, reversed (see "shuffled" label)). Behavioral measures of trained TSOM are given for different learning epochs: 5, 15, 25, 50, and 100 as the final epoch: ISR corresponds to the ability of the TSOM to correctly recall the temporal sequence of a full form at the final learning epoch (100), given its a-temporal topological representation. anticip.BMU correponds to the ability of the TSOM to predict each ensuing symbol given the preceding symbol in a full form.
