Методы компьютерной и корпусной лингвистики для решения задач эмоционального анализа интернет-текстов

Saratov : IPR MEDIA, 2022.

Kolmogorova A., Калинин А. А., Маликова А. В., Кушко Л. А.

Leading author: Kolmogorova A.

The monograph highlights the application in linguistics of a number of end-to-end technologies listed in the National Program "Digital Economy of the Russian Federation", such as big data storage and analysis technologies, artificial intelligence. Extracting opinions, detecting emotions are topics that are extremely in demand in modern research carried out at the intersection of linguistics and computer science. This publication solves a non—trivial task - to create a model that would be able to carry out multi-class attribution (eight emotional classes of texts and one neutral). In different structural parts of the monograph, discriminant features of emotional and neutral texts are described; interface design for emotional markup; cognitive strategies for emotional evaluation of texts by assessors; the values of completeness and accuracy of classification when using different machine learning models; the concept of the application, within which it is planned to implement a proven model. The work was carried out following the results of the RFBR research project dedicated to the creation of a computer classifier of Internet texts in Russian by the criterion of their emotional tonality. The monograph is intended for a wide range of specialists: linguists, specialists in the field of machine learning, psychologists, sociologists. It can be useful for students and undergraduates of enlarged groups 02.00.00 "Computer and Information Sciences", 37.00.00 "Psychological Sciences", 45.00.00 "Linguistics and literary studies"

Keywords: машинное обучение, sentiment analysis, machine learning, сентимент анализ, emotional assessment, автоматизированная оценка эмоциональных реакций