Semiotic Analysis of Texts and Interpretation of Sign Systems in the Digital Era: Sentiment-analysis Using the KNIME Platform

Introduction. The aim of the article is to study the feasibility of integrating semiotic approaches and machine learning methods for Sentiment-analysis. Sentiment-analysis is a popular area of linguistics at the interface with computer science and data analysis. The novelty of the paper lies in the attempt to interpret the results of machine learning based on the text of reviews as sign systems, revealing their lexical, syntactic, and pragmatic characteristics. Methodology and sources . The research is based on the fundamental principles of semantics, syntactics, and pragmatics, as well as on modern approaches to the automation of textual information processing and the application of mathematical methods to substantiate speech phenomena. The research material is a freely distributed data set of film reviews from the IMDB platform. The KNIME system for data analysis in the ‘No-coding’ paradigm is used as an automation tool. The paper presents a workflow including the stages of data preprocessing, construction of classification models, and evaluation of their effectiveness, and proposes a linguistic interpretation of automatic review classification errors. Results and discussion . The results demonstrate high classification accuracy (up to 92,0 %) and the ability of the algorithms to identify key lexical and syntactic markers that form the emotional colouring of the text. The study extends the boundaries of traditional semiotics by integrating methods of machine learning and big data analysis, and emphasises the practical value of using KNIME in natural language processing tasks. Conclusion. This paper provides a detailed description of an algorithm for automating Sentiment analysis of film reviews, taking into account the advantages and potential challenges of this approach for text interpretation. Prospects for further research include applying the proposed methods to multilingual corpora and analysing multimodal data, which opens up new opportunities for studying sign systems in digital communication. The proposed methodology can be applied in the commercial sphere to identify the attitudes of users to goods, services, applications, books, films, etc., which increases the interest in linguistic services, namely Sentiment analysis.