Integrated Tool Development For Text And Code Plagiarism

Today Plagiarism is most common approach taken to complete research, completing college assignments and proving innovation. Sometimes we copy others work as our innovation, for that we copy contents of their work. Real researchers and fake researchers are not identified. This problem occurs in both the cases code as well as text. To identify this problem there are many systems available on web forum but that system having some limitations like they work on either text or code. In the proposed system we are going to develop integrated tool for text and code plagiarism detection system. To solve this problem we have semantic similarity measure and TF-IDF technique (Term frequency and inverse document frequency). Further in semantic similarity measure there are two other techniques they are semantic similarity on word level and semantic similarity on document level. In TF-IDF, it finds importance of words in the form of weight after removing stop words. In preprocessing of documents we will use core NLP to extract and process sentences. In this system real time dataset is used which contains documents having text and code.

Introduction

Plagiarism and its automatic retrieval have attracted large attention from study to industry: various papers have been published on the topic, and many commercial software systems are being developed.

They may also get their code assignment from external public resources, especially the Internet. In some places, local companies may offer helping students partially or completely in those code projects. The Internet also includes several websites in which students can submit their code assignments but this may get duplicates. There are two main areas of possible plagiarism in the academia. Those are plagiarism in research papers, projects and publications. It also includes plagiarism that is especially applicable for students in the information technology majors. Now a day’s huge amount of digital information is both advantageous as well as disadvantageous too. Advantageous means that we can get each and every information on the net easily for reference and hence searching time for required information has reduced a lot.

Plagiarism has been termed as stealing, theft of concept, idea writings of other research scholar and presenting as inventor of it. In context to research today scholars are aware that plagiarism should be avoided but lack to understand what, why and how plagiarism occurs. Many successful plagiarism detection tools and software products have been developed. However, the detection of paraphrasing or confusing plagiarism remains a challenge because most of the existing tools are only able to detect copy-paste cases of plagiarism. Change in pattern of writing or language is most common technique employed for restructuring sentence to misguide scientific community and detection of search work is also major challenge.

Many Current techniques having capacity of exactly matched substring or some kinds of texture fingerprinting but that may not be sufficient as cases of rephrasing and rewording the content treated as different. Therefore, this work considers the problem of finding the suspected fragments that have the same semantics with the same/different syntax. This research work focuses on effort estimation with plagiarism analysis for research articles and code assignments. This is a research towards idea based plagiarism detection. Existing techniques focuses on keyword matching and fail to detect hidden patterns of plagiarism. Proposed research focuses on diverse patterns of plagiarism with innovative framework design.

Review of issues and struggle with academic plagiarism

In this topic we are reviewing the issues and struggle with academic plagiarism. Plagiarism is nothing but the use of ideas and research of other people without their permission, even the credit is not given to them. The few studies shown that student cheating is much widespread and is not easily recognizable. Most of the times the plagiarism is undetected. In academic plagiarism one of the article states that, to detect plagiarism, each essay must be read four times. But this only detects copying from published sources, copying from another essays is not often detectable. For this the students should be encouraged to model themselves on best thinking. Students should have to think more critically and originally. Plagiarism includes students, authors, professionals, journalists and others. Many consequences can be faced by all the students, authors, professionals and journalists. The consequences of plagiarism can be personal, professional, ethical, and legal. Plagiarism claim can cause a student to be suspended or expelled. Schools, Colleges, Universities take this plagiarism issue very seriously. For a professional person the plagiarism claim can damage their entire career. For plagiarism claim they can be fired or can be asked to step down from their current position. They will surely face the problem to find any other respectable job. The legal procedure of plagiarism can be serious issue. Copyrights laws are absolute. One cannot use another person’s material without citation and references. Some plagiarism issues can be a criminal offence and it may lead to prison sentence. The consequences of plagiarism have been widely reported in the world of academia. Once caught with plagiarism claim an academic’s career can be ruined. Publication is most integral part of academic career. To lose the ability to publish most likely means the end of an academic position and a destroyed reputation.

The wide availability of the Internet and computers, and the ease, with which electronic texts can be copied, are among the most obvious factors that contribute to this problem. Academic dishonesty has been a persistent part of the higher education landscape. Understanding the potential causes and complexities of academic dishonesty is critical in building an effective academic culture and system to try to counter this phenomenon.

System Architecture

Figure Shows detailed flow of Plagiarism Detecting System. In this user can upload text and code document as an input. Using core NLP technique, given text file or code file will be processed. We are going to perform operation like stemming, stop words removal and parsing technique. After this semantic similarity check will be performed on word or concept level as well as document or text level. Based on the similarity check TF-IDF values will be calculated of words present in already uploaded document. Based on TF-IDF values plagiarism report will be returned to the user in the form of duplicate content and graphical representation.

Semantic Similarity Technique

The concept of semantic similarity is fundamental and widely understood in many domains of natural language processing. It can be defined as the degree of taxonomic proximity between terms.

In the case of plagiarism detection, semantic similarity should be expressed on the text level as the final result. The score of semantic similarity between suspected document and one or more other documents may indicate the existence of plagiarism.

The whole procedure of plagiarism identification, semantic similarity can be calculated on the sentence and paragraph level.

Related works

Plagiarism seeds identification for the high-obfuscation proposed by Leilei Kong in which presents a multi-features fusion method. From suspicious document and source document, integrated lexicon features, syntax features, semantics features and structure features are extracted using this method. A multi-feature fusion classifier based on Logical Regression model is proposed to decide whether a text fragment pair can be regarded as plagiarism seeds or not. Haoliang et. al. proposed an effective method in which high-obfuscation plagiarism seeds presents a significant research problem in the field of plagiarism detection. To capture plagiarism seeds the conventional methods of plagiarism detection are used based on single type of features. Rada Mihalcea et. al. uses a methods like measuring the semantic similarity of texts and using corpus-based knowledge-based similarity measures.

S. Santhosini devi Proposed a system in which presents a measure of semantic similarity in an is a taxonomy based on the notion of shared information content. Experimental evaluation against a benchmark set of human similarity judgments demonstrates that the measure performs better than the traditional edge-counting approach.

Alexander Maedche proposed a system in which Ontology serve as a means for communication at a similarity and semantic level of the text contents. Samuel Fernando proposed a system in which presents a novel technique to the problem of paragraph identification. Although paraphrases often make use of identified or near words, many previous approaches have either ignored or made limited use of information about similarities between word meanings. James O’Shea proposed a system in which describes a comparative study of STASIS and LSA. These measures of semantic similarity can be applied to short texts for use in Conversational Agents (CAs). CAs are computer programs that interact with humans through natural language dialogue. Taras Finikov proposed a system in which influence of transformation processes in higher education to lower academic standards, changes and deformation in ethical field of global and national higher education. We considered the genesis and modern standards of academic integrity.

Conclusion

In this paper we are describing our preliminary work on semantic similarity measures and their possible usage for content detection in the task of plagiarism identification. This identification develop by TFIDF technique for both code and documents, TFIDF technique finds weight of words after removing stop words present in documents and finally generate percentage report in the form of file and graph.