Developing and Applying a Neural Network System for Text Plagiarism Detection in Higher Education

This article focuses on the development and features of applying a neural network system for text plagiarism detection in higher education. In the context of the rapid growth in the availability of open information resources, the problem of academic honesty is becoming increasingly important. This article presents an approach that uses machine learning techniques, specifically a fully-connected neural network, to automatically detect plagiarism in academic and research works. The main stages of dataset creation, model development, and training are discussed, and the results of applying the model in practice are analyzed. Based on the test results, the average absolute error for the architecture with two hidden layers was found to be 9.38. The results of the study demonstrate the effectiveness of the proposed system in identifying plagiarism, making it a valuable tool to promote academic integrity and ensure educational quality.