Learning to pronounce written words : a study in inductive language learning

ion by compression, 23, 144 levels of, 10, 12, 16, 46, 47 activation, 29 propagation, 29 alphabet orthographic, 179 phonemic, 179 alphabetic writing system, 2 analogical modelling, 10 analogy, 2 analogy principle, 2 arbiter model, 151 module, 148, 151 arcing, 150 ART/S, 113–116 experiments with, 115 articulatory features, 48, 180 artificial data, 155 artificial neural networks, 28 associative relation, 7 back-propagation, 28–30, 185 bagging, 150 benchmark problem, 166 bias class, 56 boosting, 150 BP, 28, 140 C4.5, 35, 39–41, 142 case-based reasoning, 21 CELEX, 26, 50 Charivarius, 14 chunk-based word pronunciation, 167 classification, 8 cluster in instance space, 156 clustering, 13 co-articulation, 44 coda, 49 combiner model, 150 module, 147, 150 compounding, 44 compression, 23, 24, 37, 38 connection, 29 weight, 29 connectionism, 28 consonant, 49 corpus-based linguistics, 2 correspondence, 2 data characteristics, 153–165 DC, 56 decision tree learning, 39 decision trees, 37 induction of, 37