ordbogsbaserede

Ordbogsbaserede refers to methods, systems or resources that rely primarily on dictionaries (ordbøger) or lexical databases as the main source of linguistic information. In practice, a dictionary-based approach uses lexicon entries for words, including lemma, part of speech, inflection, sense distinctions and sometimes lexical collocations. Such approaches are common in natural language processing tasks such as lemmatization, morphological analysis, spell checking and initial word sense disambiguation, often supplemented by rule-based modules to handle morphology and syntax.

In spell checking, the algorithm compares surface forms against a dictionary and uses morphological rules to validate or suggest corrections. In morphological analysers, the lexicon provides base forms and inflection patterns, enabling the generation of lemmas from inflected forms. In word sense disambiguation, dictionary senses provide candidate meanings; context-based scoring can be applied, though disambiguation is limited by the richness of the dictionary's sense inventory.

Advantages include precision for vocabulary covered by the lexicon, fast lookup and interpretability. Limitations include dependence on dictionary coverage, difficulty handling neologisms, proper nouns, slang or multiword expressions not present in the dictionary; updating dictionaries is necessary to maintain coverage; the approach is less effective for languages with sparse lexical resources.

In practice, ordbogsbaserede methods are often used in hybrid systems that combine dictionary information with corpus-based statistics or machine learning, leveraging the reliability of lexicons while benefiting from contextual data.