Studying Comments on Russian Patriotic Actions: Sentiment Analysis Using NLP Techniques and ML Approaches
Anton Sergeevich Sysoev, Андрей Александрович Линченко, V. A. Kalitvin, Daniil Aleksandrovich Anikin, Oksana Vladimirovna GOLOVASHINA · 2021 3rd International Conference on Control Systems, Mathematical Modeling, Automation and Energy Efficiency (SUMMA) · 2021
Due to the increasing number of online social media resources and social networks the problem of natural language processing (NLP) applying machine learning (ML) approaches is becoming an important concept in monitoring non-formalised text data and catching information in real time. Studies in historical memory and appropriate patriotic actions have to be based on detailed people’s moods examination. The presented study contains approaches to analyse the polarity of comments on patriotic actions in Russia. Described scheme of NLP text processing and ML classifier were used in the analysis of two regions. There were used standard approaches to convert natural texts into numerical representation. To identify the polarity of the text several models including linear and deep neural networks were examined. Long-short-term-memory network has demonstrated the best result on validation set and was used to the solve the stated problem.