A Machine learning–enabled framework for measuring and evaluating brand equity in the telecommunications sector

This study was conducted with the aim of developing a data-driven model for assessing and analyzing brand equity in the telecommunications industry. In the first stage, the factors influencing brand equity were identified through a literature review and expert surveys, and then weighted using the Fuzzy Best–Worst Method (FBWM) to determine the relative importance of each dimension. Subsequently, data collected from a structured questionnaire consisting of 75 indicators and 1,980 valid records were used to develop the machine learning model. At this stage, a multinomial logistic regression algorithm was employed to classify brand equity into three levels: “weak,” “moderate,” and “strong.” The analysis results revealed that the dimensions of conformity and loyalty, brand image, brand performance, and brand trust played the most significant roles in shaping overall brand equity. The final model achieved an accuracy rate of 81 percent, demonstrating strong performance in distinguishing the two extreme classes (weak and strong). The findings indicate that in the digital services sector, brand equity is not solely dependent on technical quality but emerges from the interaction among trust, emotional experience, behavioral loyalty, and a positive brand image. By integrating a multi-criteria decision-making approach with machine learning techniques, the proposed model provides an effective analytical tool for brand data evaluation and supports decision-making in marketing strategy and customer experience management.