Subscribe to the gold package and get unlimited access to Shamra Academy
Register a new userEnglish-Arabic Cross-language Plagiarism Detection
انتشار الانتحال باللغة الإنجليزية والعربية
337
0 0
0.0
(
0
)
Created by
Shamra Editor
Ask ChatGPT about the research
ساهم تقدم تقنية الويب والمعلومات في النمو السريع للمكتبات الرقمية وأدوات ترجمة الآلات التلقائية والتي تقوم بسهولة بترجمة النصوص من لغة إلى أخرى. وقد زادت هذه المحتوى في الوصول إلى لغات مختلفة، مما يؤدي إلى أداء الانتحال المترجم بسهولة، يشار إليها باسم الانتحال عبر اللغة ". التعرف على الانتحال بين النصوص بلغات مختلفة هو أكثر تحديا من تحديد الانتحال داخل وجعة مكتوبة بنفس اللغة. تقترح هذه الورقة تقنية جديدة لتعزيز اكتشاف الانتحال باللغة الإنجليزية والعربية على مستوى الجملة. تستند هذه التقنية إلى استخراج ميزة دلالية ونقصية باستخدام ترتيب Word و Word AdgetDing و MIGNIMENT مع ترميزات متعددة اللغات. يتم بعد ذلك استخدام هذه الميزات ومجمديها مع خوارزميات مختلفة لتعلم الآلات (ML) من أجل المساعدة في تصنيف الجمل كإخلاء إما مكسوين أو غير متسائل. تم نشر النهج المقترح وتقييمه باستخدام مجموعات البيانات المقدمة في Semeval-2017. يوضح تحليل البيانات التجريبية المستخدمة في استخدام الميزات المستخرجة ومجموعاتها مع مختلف منصات ML، تحقق نتائج واعدة.
The advancement of the web and information technology has contributed to the rapid growth of digital libraries and automatic machine translation tools which easily translate texts from one language into another. These have increased the content accessible in different languages, which results in easily performing translated plagiarism, which are referred to as cross-language plagiarism''. Recognition of plagiarism among texts in different languages is more challenging than identifying plagiarism within a corpus written in the same language. This paper proposes a new technique for enhancing English-Arabic cross-language plagiarism detection at the sentence level. This technique is based on semantic and syntactic feature extraction using word order, word embedding and word alignment with multilingual encoders. Those features, and their combination with different machine learning (ML) algorithms, are then used in order to aid the task of classifying sentences as either plagiarized or non-plagiarized. The proposed approach has been deployed and assessed using datasets presented at SemEval-2017. Analysis of experimental data demonstrates that utilizing extracted features and their combinations with various ML classifiers achieves promising results.
References used
https://aclanthology.org/
rate research
Read More
This paper presents a review of available algorithms and plagiarism detection systems، and an
implementation of Plagiarism Detection System using available search engines on the web.
Plagiarism detection in natural language documents is a complicat
ed problem and it is related to the
characteristics of the language itself.
There are many available algorithms for plagiarism detection in natural languages .Generally these
algorithms belong to two main categories ; the first one is plagiarism detection algorithms based on
fingerprint and the second is plagiarism detection algorithms based on content comparison and includes
string matching and tree matching algorithms .
Usually available systems of plagiarism detection use specific type of detection algorithms or use a
mixture of detection algorithms to achieve effective detection systems (fast and accurate).
In this research, a plagiarism detection system has been developed using Bing search engine and a
plagiarism detection algorithm based on Rhetorical Structure Theory.
We introduce HateBERT, a re-trained BERT model for abusive language detection in English. The model was trained on RAL-E, a large-scale dataset of Reddit comments in English from communities banned for being offensive, abusive, or hateful that we hav
e curated and made available to the public. We present the results of a detailed comparison between a general pre-trained language model and the retrained version on three English datasets for offensive, abusive language and hate speech detection tasks. In all datasets, HateBERT outperforms the corresponding general BERT model. We also discuss a battery of experiments comparing the portability of the fine-tuned models across the datasets, suggesting that portability is affected by compatibility of the annotated phenomena.
In this paper we review and list, the advantages and limitations of the significant
effective techniques employed or developed in text plagiarism detection. It was found that
many of the proposed methods for plagiarism detection have a weakness poi
nts and do not
detect some types of plagiarized operations.
This paper show a survey about plagiarism detection including several important
subjects in plagiarism detection, which is plagiarism definition, plagiarism prevention and
detection, plagiarism detection systems, plagiarism detection processes and some of the
current plagiarism detection techniques.
This paper compares between different plagiarism detection algorithms, and shows
the points of weakness, and points of efficiency, and describe the power of semantic
plagiarism detection methods, and shows its efficiency in detect plagiarism cases that
another plagiarism detection algorithms don’t able to detect these cases, that semantic
plagiarism detection methods are developed to get rid of traditional weakness points for all
plagiarism detection methods have.
This paper presents a reference study of available algorithms for plagiarism
detection and it develops semantic plagiarism detection algorithm for plagiarism detection
in medical research papers by employing the Medical Ontologies available on the
World
Wide Web.
The issue of plagiarism detection in medical research written in natural languages is
a complex issue and related exact domain of medical research.
There are many used algorithms for plagiarism detection in natural language, which
are generally divided into two main categories, the first one is comparison algorithms
between files by using fingerprints of files, and files content comparison algorithms, which
include strings matching algorithms and text and tree matching algorithms.
Recently a lot of research in the field of semantic plagiarism detection algorithms
and semantic plagiarism detection algorithms were developed basing of citation analysis
models in scientific research.
In this research a system for plagiarism detection was developed using “Bing” search
engine, where tow type of ontologies used in this system, public ontology as wordNet and
many standard international ontologies in medical domain as Diseases ontology which
contains a descriptions about diseases and definitions of it and the derivation between
diseases.
This paper deals with automatic detection of plagiarism in Arabic documents. We present in this paper a new idea based on the experimentation of lexical chains. The proposed method extracts those chains from original document and uses a search engine
to verify if such chains occur in other documents. The second step in our methods uses automatic translation system to translate lexical chains and verify by using search engine if those chain occurs in document in other languages. Then we compute a correlation ratio between lexical chains and lexical chains extracted from documents provided by the search engine to detect plagiarism in the original document.
We present in the end of this paper our prototype called « Alkachef » developed to detect plagiarism in Arabic document .
suggested questions
Log in to be able to interact and post comments
comments
Fetching comments
Sign in to be able to follow your search criteria
