Skip to content

Incorrect language is being returned for specific words #57

Description

@simoncast

Hi,

The words "System Scaling Strategy" will return anything but English it seems. I've had West Frisian, Afrikaans and now Swedish as the "best result".

This is the output of language detection:
( 'sv' => 0.4363359707851491, 'nl' => 0.43365794278758374, 'hu' => 0.4318320146074255, 'nb' => 0.42452830188679247, 'de' => 0.4225197808886184, 'da' => 0.4211199026171637, 'en' => 0.41880706025562997, 'la' => 0.4111990261716373, 'pt-BR' => 0.4084601339013999, 'pt-PT' => 0.40833840535605603, 'id' => 0.40626902008520993, 'is' => 0.3883749239196591, 'lt' => 0.38758368837492396, 'sl' => 0.38533171028606206, 'ro' => 0.3756542909312234, 'es' => 0.37321972002434567, 'it' => 0.3685940353012781, 'et' => 0.3680462568472307, 'cs' => 0.3612903225806452, 'cy' => 0.3496652465003043, 'pl' => 0.3467437614120511, 'bs-Latn' => 0.3446743761412051, 'lv' => 0.3444917833231893, 'fr' => 0.3375532562385879, 'tr' => 0.33609251369446136, 'hr' => 0.33049300060864273, 'sk' => 0.32860620815581254, 'fi' => 0.3032258064516129, 'gd' => 0.29519172245891656, 'ga' => 0.28387096774193543, 'ka' => 0, 'ml' => 0, 'hy' => 0, 'el-polyton' => 0, 'el-monoton' => 0, 'uk' => 0, 'zh-Hans' => 0, 'zh-Hant' => 0, )

I limited the languages to exclude Frisian and Afrikaans. Why is English not the first in the list?

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions