DOI: 10.17587/prin.11.54-64

Software System for Stance Detection Based on Compositional Approach

S. V. Vychegzhanin , vychegzhaninsv@gmail.com, Vyatka State University, Kirov, 610000, Russian Federation

Corresponding author: Vychegzhanin Sergey V., Software Engineer, Vyatka State University, Kirov, 610000, Russian Federation, E-mail: vychegzhaninsv@gmail.com

The article is devoted to the task of automatic stance detection. Stance detection is the task of automatically determining from text whether the author of the text is in favor of, against, or neutral towards a proposition or target. There is a wide range of areas where stance detection is used, including political, sociological and marketing research, the search engines, the human-computer interfaces. The article proposes the structure of the software system for stance detection based on ensembles of methods for feature selection and ensembles of classifiers. The structure includes four main subsystems, namely text preprocessing, feature space dimension reduction, creation of ensembles of classifiers and analysis subsystems. The system is implemented in accordance with the object-oriented approach. Main classes and their relations are described with UML class diagram. The results of experimental research of developed system are presented. The research is performed using the text corpora composed of messages from users of the social network "VK" and online forums. These results show high quality stance detection that is superior to other machine learning methods. The developed system can be used as a standalone application for stance detection. It is possible to integrate the system into third party services, that search and recommend documents based on information analysis.

Keywords: natural language processing, opinion mining, stance detection, software system, class diagram

pp. 54–64

For citation:
Vychegzhanin S. V. Software System for Stance Detection Based on Compositional Approach, Programmnaya Ingeneria,2020, vol. 11, no. 1, pp. 54—64

This work was supported by the Ministry of Education and Science of the Russian Federation and by the Vyatka State University (the project "Development and research of sentiment lexicons for text sentiment analysis" No. 34.2092.2017/ 4.6).