Institutional Repository of Vilnius University
Natūralios kalbos apdorojimo ir sentimentų analizės metodų taikymas įmonių reputacijai vertinti /
Abstract
dc:descriptionInternet technologies are increasingly prevalent in everyday life, encompassing social networks, online stores, and news portals. These developments generate a growing flow of data, generated by internet users, including comments, feedback, articles, and discussions. Appropriately utilized, such data can create significant added value, such as when an organization releases a new product and monitors user feedback online. However, this monitoring process presents two challenges: selecting relevant texts and assessing their potential impact on the organization's reputation. This study addresses a two-part task by employing natural language processing methods. Models for entity recognition are utilized to select texts related to the organization. The methodology for employing and evaluating these models is defined, including discussions on training data sets and assessing the suitability of models for the selected research data. Finally, conclusions are drawn regarding the most suitable model for solving this task. Another aspect of the study involves evaluating the impact of selected texts on the organization's reputation. An expert dataset on reputation management systems, compiled by experts, is employed. Differences between text sentiment values and potential textual influence on the organization's reputation are discussed. These differences are assessed using a transformer model trained for sentiment analysis. The investigation into textual influence on reputation concludes with training several different transformer models and evaluating the classification results of these models. In conclusion, the study proposes the most fitting models to meet the requirements, evaluates further research possibilities, and presents key findings that emerged throughout the course of the study.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2023
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Pranauskas, Gvidas,
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD192845412&prefLang=en_US