Application of data warehouses for plagiarism detection in text documents

The work is devoted to the application of data warehouses for detecting plagiarism in text documents, as a result of which a scalable software architecture is built. A dataset of bachelor's and master's theses was prepared and pre-processed, including cleaning, lemmatization, removal of stop words, removal of symbol substitutions, replacement with synonyms. A test dataset was created to analyze the performance of the solution. A study of the efficiency of the applied data stores was conducted, their comparative analysis was performed, and several runs were conducted to avoid stochastic error. Based on the results of model testing, DynamoDB was singled out as the most efficient data warehouse for the plagiarism detection task. Bibl. 14, fig. 8, tabl. 3