TY - JOUR

T1 - Social Media Sentiment Analysis with Context Space Model

AU - Мальцева, Анна Васильевна

AU - Шилкина, Наталья Егоровна

AU - Махныткина, Олеся Владимировна

AU - Лизунова, Инна

N1 - Publisher Copyright: © Springer Nature Switzerland AG 2020.

PY - 2020/1/1

Y1 - 2020/1/1

N2 - In this article the description of algorithm of an assessment of mood of the statement is presented with the accent on the context of user’s messages in social media. The article focuses on the fact that messages containing identical sentiment objects have different meaning that affects onto the evaluation of the sentiment of the message. An additional research objective is the identification of formal criteria for assigning messages to classes “core”, “periphery”, “non-relevant” to denote the role of the research relevance of the object key in the message. In this article, we have given several examples of authentic messages for each group. The method was tested on the empirical basis of more than 10,000 messages to assess the relationship of users of the social network VKontakte to the object of tonality – a form of employment “freelance”. The research methodology presupposes the use of basic and additional methods of data preprocessing, data augmentation, comparative analysis of the application of classification methods. The article includes comparative description of results of application logistic regression, support vector machines, naive Bayesian classifier, nearest neighbor, random forest.

AB - In this article the description of algorithm of an assessment of mood of the statement is presented with the accent on the context of user’s messages in social media. The article focuses on the fact that messages containing identical sentiment objects have different meaning that affects onto the evaluation of the sentiment of the message. An additional research objective is the identification of formal criteria for assigning messages to classes “core”, “periphery”, “non-relevant” to denote the role of the research relevance of the object key in the message. In this article, we have given several examples of authentic messages for each group. The method was tested on the empirical basis of more than 10,000 messages to assess the relationship of users of the social network VKontakte to the object of tonality – a form of employment “freelance”. The research methodology presupposes the use of basic and additional methods of data preprocessing, data augmentation, comparative analysis of the application of classification methods. The article includes comparative description of results of application logistic regression, support vector machines, naive Bayesian classifier, nearest neighbor, random forest.

KW - Machine learning

KW - Sentiment-analysis

KW - Social media sites

KW - Tonality

UR - http://www.scopus.com/inward/record.url?scp=85082558396&partnerID=8YFLogxK

UR - https://www.mendeley.com/catalogue/32e577aa-c735-340d-9e92-d2dbc97a7b6d/

U2 - 10.1007/978-3-030-39296-3_29

DO - 10.1007/978-3-030-39296-3_29

M3 - Article

VL - 1135 CCIS

SP - 399

EP - 412

JO - Communications in Computer and Information Science

JF - Communications in Computer and Information Science

SN - 1865-0929

ER -