Browse Prior Art Database

Recommending and Ranking Foreign Documents Based on their Difficulty

IP.com Disclosure Number: IPCOM000199805D
Publication Date: 2010-Sep-16
Document File: 5 page(s) / 88K

Publishing Venue

The IP.com Prior Art Database

Abstract

A computational method is provided for retrieving documents which are not only relevant to the query provided by the user, but also match the user's degree of comprehension of foreign languages.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 54% of the total text.

Page 1 of 5

Recommending and Ranking Foreign Documents Based on their Difficulty

Disclosed is a computational method for retrieving documents which are not only relevant to the query provided by the user, but also match the user's degree of comprehension of foreignlanguages.

Consider that the user of a information retrieval system is learning a foreign language. How can the user find a foreign book that properly matches his/her language understanding, ie, that closely approximates his/her language comprehension level?

Moreover, consider that the user is searching multiple documents online which may be in many foreign languages. Different documents may answer to a different extent the user's query.

Additionally, a document's difficulty or understanding based on the language it is

written, may increase or decrease its difficulty. How to order them from least "difficult" (or readable) to most "difficult" but also consider the

document's topic relevance?

Such operations are essential nowadays that many documents exist in different languages (and users arebecoming more accustomed

with multilingual corpora).

How can the user be provided with the most relevantand at the same time readable document from a set of appropriate multilingual documents,

while considering

both its relevance and potential level of understanding? The method here described provides the solution to this problem.

Relevant to this patent are "readability" metrics for same language corpora,

which are significantly simpler, because

they only take into consideration features such as number of words per sentence or sentence length.

When considering foreign language documents, it theoriginal language of the user should also be considered in order to identify 'cognates ',

words that may be similar in the two languages and therefore enhance the reader's understanding of a document.

Additionally, the overall difficulty of a foreign document must

readability techniques.

APPLICATIONS:

Language-Comprehension personalization of the Web. Using our methodology one can can rank and present 'similar' news articles to a foreign language user, based on the perceived comprehension of the article, that is, from most easy to most advanced usage of the foreign language.

Augmented with a

we envision a

An approach

such the one presented in this work, presents a solution to this problem.

be assessed, something that is also not consideredin single language

pre-trained portion that stored the language level of the user,

multilingual personalization of the web.

Online Bookstores. Imagine the case of a English speaking reader that is interested in reading German literature books. Which one should he/she read based on one's level of foreign language comprehension?

1

Page 2 of 5

With the advent of online bookstore places in new many mobile devices,

multiple cases applicable to this invention.

REFERENCES:

[1] Predicting reading difficulty with statistical language models Kevyn Collins-Thompson, Jamie Callan

htt...