Сравнительный анализ методов машинного обучения для решения задачи классификации документов научно-образовательного учреждения

Abstract

This article discusses the actual problem of classification of documents using machine learning methods in the subject area of research and educational institutions. Analysis of developments in this area showed that there is no sufficient theoretical basis for the integration of existing classification methods for the analysis of documents of research and educational institutions. Therefore, to solve this problem, an algorithm of classification of documents, taking into account the specifics of the documents of the subject area of scientific and educational institutions. The article deals with the system of features used to solve the problem of combined classification. The paper considers the approach of preprocessing of the text, which allows using the known methods of machine learning to improve the accuracy and speed of document clas-sification.