Skip to content

Add word part frequencies to decompounder tree #10

@HannaLindgren

Description

@HannaLindgren

By adding a frequency number to the compounding word parts, this can be used to select a best guess when several alternatives with the same number of compound parts are found.

Furthermore, very high frequent suffixes can be used for determining compound boundaries, although the prefix is not known (e.g., in Swedish "-gata" (street) may be such a safe bet).

[moved here from Phabricator Wikispeech]

Metadata

Metadata

Assignees

No one assigned

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions