Weighted semantic plagiarism detection approach based on AHP decision model.

Area of Science:

• Computer Science
• Information Science

Background:

• Academic plagiarism is a growing concern, with sophisticated methods used to evade detection systems.
• Existing plagiarism detection algorithms often struggle with semantic variations and inverted word order, and may not optimize for time complexity.

Purpose of the Study:

• To propose an effective method for detecting structural and semantic plagiarism.
• To reduce the time complexity of plagiarism detection by analyzing only essential parts of documents.
• To improve the accuracy of plagiarism detection, even when word order is altered.

Main Methods:

• Utilizing a subset of document content for similarity calculation, rather than the entire text.
• Employing a set of significant terms and their combinations to enhance detection capabilities.
• Assigning differential weights to words based on their location within document sections.
• Implementing an Analytical Hierarchy Process (AHP) model for weighted similarity calculation.

Main Results:

• The proposed approach demonstrates superior ability in detecting semantic academic plagiarism.
• A significant reduction in runtime was observed compared to existing plagiarism detection methods.
• The method effectively handles variations in word order and synonym usage.

Conclusions:

• The developed method offers an efficient and accurate solution for academic plagiarism detection.
• By focusing on key terms and positional weighting, the system overcomes limitations of traditional approaches.
• This research contributes to more robust and time-efficient academic integrity tools.