ComputingCommunicationComputer ScienceComputational LinguisticsNatural Language ProcessingText AnalyticsLanguage ModelingLanguage ModelsN-gram AnalysisSyntactic AnalysisText GenerationWord OrderAlternative Communication DevicesAlternative Communication InterventionLexical PredictionPredictionSyntactic PredictionSyntactic RepresentationText-based AacWord N-grams
Most icon-based augmentative and alternative communication (AAC) devices require users to formulate messages in syntactic order in order to produce syntactic utterances. Reliance on syntactic ordering, however, may not be appropriate for individuals with limited or emerging literacy skills. Some of these users may benefit from unordered message formulation accompanied by automatic message expansion to generate syntactically correct messages. Facilitating communication via unordered message formulation, however, requires new methods of prediction. This paper describes a novel approach to word prediction using semantic grams, or “sem-grams,” which provide relational information about message components regardless of word order. Performance of four word-level prediction algorithms, two based on sem-grams and two based on n-grams, were compared on a corpus of informal blogs. Results showed that sem-grams yield accurate word prediction, but lack prediction coverage. Hybrid methods that combine n-gram and sem-gram approaches may be viable for unordered prediction in AAC. 1
30
Effects of Age and Gender on Blogging
Jonathan Schler, Moshe Koppel, Shlomo Argamon et al. · National Conference on Artificial Intelligence · 2005 · 558 citations
Unsupervised Multilingual Sentence Boundary Detection
Tibor Kiss, Jan Strunk · Computational Linguistics · 2006 · 421 citations · Full text
Positional language models for information retrieval
Yuanhua Lv, ChengXiang Zhai · 2009 · 230 citations