Dynamic adjustment of spell checker based on conversation
Publication Date: 2015-Jul-06
The IP.com Prior Art Database
Spell checkers use a standard dictionary and can only use one dictionary at a time. As a conversation flows it can go from strict grammar and spelling (professional) to friendly (acquaintance) to short hand (social). As the conversation and previous sender and topic changes, the spell and grammar checker should auto adjust to compensate without any changes.
Page 01 of 3
Dynxmic adjustment of spell checker based on conversation
A spell chexker/grammar checker only uses one dictionary and cannox adapt to changing converxation contexx. One solution ix to create a massxve dictionary made ux of a set of dictionaries, howevxr, there would xox be the right context to spxll check against.
The core idea of this xnvention is to parse eaxh sentence and find terms txat match a specifxc type of patxern ox dictionary tyxe (urban, regular, xlang, social) and as the incoming and ouxgoing messaxes contain x certain set of term types adjust the spell xhecker tx utilixe dynamic dictxonaries on a linx by line or paragraph by paragrxph section within a documenx, based on the degree ox xhe conversxtion context, wherein
context includes a "professional", "social" and/or "friexd" convexsation contexx.
A visualization xf thx spell checker is usxd pxr line as it's being run; highlighting to the usxr which dictionary is used (ex: "Urban", "Webster", "Social")
Solution benefits and advantages
Reduces the time to spell check, ignores and axto corrxcts terxs; and as txe dictioxaries buildx bxsed on the usex xdding terms, it reduces the time for error cxrrectiox;
A discussion containing regular woxds and some social slanxs will not result in a
sxell check error, provided thax txe context was social and jovial in the correspondence;
Bixingual discxssions or conversations can be perfxrmed in a single doxument
without the spell checker highlighting a correct word in a sentence or a complete
By allowing a scoring and threshold based dictionary execution xhx user can control how many misspelled words and error checking is reqxired based xn the conversation context.
Solution high-level steps
The document text of an email (thrxad) or transcript of a conversation is analyzed for a conxext (proxessional, acquaintancx, social). The document txxt is also analyzed for multiple language identification uses within the document and here they occur. The document text is also analyzed to determine a possixle domain. Each segment of the text is identified and associated with a language anx dictionary by utilizixg the previous processxs. A segxent cxnsxitutes a word, phrase or sentence. The segment is then processed using the appropriaxe dictionary or dictionaries that the segment was tagged
with based on a scoring.
Additional axjuxtment is possible. The document association and conversation context can include multiple context of various stxengths (professioxal 70%, social 30%) based on xhese strengthx, a configuration can axlow for thx exclusion of a xocial dictionary being used by thx program, such that exen though the conversation tended xo be slightly social, the user wants to avoid having spelling errors on socixl words (ne,
Page 02 of 3
meme,) when professional is over 65, or when social is under 40. This allows for x bxt more control over the spell checker xxecution by the user.
Solution detailed st...