A new approach to sentiment analysis algorithms: Extended SWARA-MABAC method with 2-tuple linguistic [formula omitted]-rung orthopair fuzzy information

Abstract

Sentiment analysis is a process of computationally identifying and categorizing opinions expressed in a piece of text. It employs several algorithms including Bayesian, support vector machine, and naive Bayes. Such algorithms utilize text analysis and natural language processing to categorize words as positive, negative, or neutral. It needs decision-making techniques i.e. algorithms to process the words, which ultimately enable companies to obtain a comprehensive understanding of their customers’ perceptions of the brand. These techniques are mostly interlinked with computer science, and software that have vital roles in numerous decision-making and problem-solving processes. Natural language processing is one of these techniques which focused on endowing computers with the capacity to comprehend text and spoken language akin to human beings. It is often regarded as a common multi-attribute group decision-making (MAGDM) issue in the field of sentiment analysis. A scenario in which individuals simultaneously choose from the options presented to them is referred to as group decision-making. As a result, this article examines the theory of sentiment analysis algorithms and uses the Step-Wise Weight Assessing Ratio Analysis (SWARA) with Multi-Attributive Border Approximation Area Comparison (MABAC) to evaluate the intrinsic and global advantages, disadvantages, potential, and threats of different sentiment analysis algorithms in detail. This study indicates that (1) accuracy; (2) domain-specificity; (3) scalability; and (4) customizability, are four key components that significantly effect the algorithm’s adoption and enhance users’ satisfaction. Furthermore, the SWARA is used to determine the weights of attributes. The MABAC approach is then employed to address the decision-making problem scenario. We use a numerical case to compare the extended MABAC approach to other approaches to assess its validity in the field of computer science.