naapurisanat

Naapurisanat, or neighboring words, is a concept used in Finnish linguistics and corpus linguistics to describe the words that appear in close textual proximity to a target word. They constitute the surrounding context within a defined window and are used to infer meaning, syntax, and usage patterns.

In practice, researchers collect naapurisanat by examining a fixed context window around each target word. The window can be measured in number of words (for example, two or five on each side) or in syntactic units. Neighbors may be left, right, or both, and are often tokenized and lemmatized to accommodate Finnish inflection.

Naapurisanat are central to co-occurrence analysis and to methods that generate word embeddings. They help reveal collocations, semantic fields, and typical syntactic constructions. The choice of preprocessing, such as handling punctuation and sentence boundaries, influences which words count as neighbors.

Applications include lexicography, language teaching, information retrieval, and natural language processing tasks such as word sense disambiguation and translation. By studying naapurisanat, lexicographers and developers gain insight into how words co-behave in real text.

Example: In the sentence "Kissa istuu pöydällä ja katselee kärpäsiä," the naapurisanat for "kissa" within a two-word window are "istuu" and "pöydällä."

See also: collocation, co-occurrence, context window, word embedding.