LFC - Language from context

Objective

help people learn new languages.

Steps:

determine language proficiency
send daily email with short text, tailored to the proficiency

Language proficiency Assessment

Via a quiz or a short form where the user indicates which words are known and which are unknown, we can make a mental model of what are the things the student knows and what they don't.

This can be a long form of words or we may have to figure out a better way of assessing knowledge of the language.

Daily Email Delivery

The system sends email to the users with the words the system assumes are known being 95% of the contents of the email and 5% of words which are supposed to be unknown.

The unknown words can be added in a short vocabulary list at the bottom of the email. Including their definition in the target language and the translation to the student's language

The system selects articles or paragraphs from a list of existing content or perhaps news content, this content is used as a source of information, and the system, selects only content that matches the criteria. The system needs a word index that can match a given article to a list of words, this word index contains a connection between articles and words, and after the index is created, the system finds articles where 95% of the words in the article are known to a given student.

Each student is a record on the database, and for each student, a list of known words are stored.

The words are also ranked by how common they are, so that more common words are put on top of the results.

This is so that the student learns the most common words first, and later proceeds to learn less common words.

Assuming enough content is found, the system could categorize the source article database and find the correct articles in each case.

The email contains a link to a open platform where the user can indicate which words are known and which words are not known, so the users can train the system constantly regarding newly acquired words that may have originated from external sources during the process.

The user interface is simple, and lets users indicate words by clicking on words they don't know.

System Entities

Language

languageId
languageName

Word

languageId
WordId
word
index
category (grammar, vocab, etc).

Translation

sourceLanguageId
destinationLanguageId
sourceWordId
translationText

Definition

languageId
wordId
definitionText

Article

ArticleId
title
contents
length (words)

wordsDefinedBy() (This returns which words are defined by this article)

ArticleDefinition

articleId
wordId

WordIndex

wordId
articleId

buildIndex() (build an index of all words from all articles for fast checking)

Student

studentId
Name
level (begginner, intermediate, advanced=
Email
FBID
desiredKGrowth
active
sourceLanguageId

sendMessage(title, body)
getAllStudents()
createNewStudent(name,level,email,fbid,kgrowth)
unsuscribeStudent(email) (users should be able to unsuscribe at any time)
setDesiredGrowth(email, growth) users can indicate how /fast/ they want to learn.
markWordAsKnown(email, word) the user clicks on word to indicate they know them as a result the system will assume knowledge of new words these words may have come from the list of new vocabulary or from any where in the article.
getGrowthGraph(email, dateStart,dateEnd) the system can know how many new words are being acquired by a student in a given period.

Target

targetId
studentId
languageId
sourceLanguageId

suscribeToDailyMessage(source, dest, email)

Knowledge

studentId
wordId
dateLearned

showStudentKnowledgeForm() (shows a form to a student so he can indicate what he knows and what he doesn't)

Message

messageId
languageId
studentId
timestamp
title
body
kgrowth

calculateKGrowth() calculates how many words are actually unknown in the message and returns a number from 0% to 100%
getMessageVocabularyTable(title, body) builds a vocabulary table of words which are supposed to be unknown to the student from the contents of the article and uses as a
sendAllMessages() sends a message to all students, this can be run as a cron job.
sendMonthlyReport() sends a montly report to all students with knowledge acquisition information so they can visualize the progress email includes a graph and a list of words.
getMessageTemplate() returns the template for the email format
showMessageOnline(messageId) shows a message online so students can click on it

adammendoza / language-from-context Goto Github PK

language-from-context's Introduction