Comparative analysis of fresh food e-commerce brand attitudes based on STM theme model

The circulation of fresh agricultural products is related to the quality of consumption and agricultural development. This article takes the online reviews of fresh products as the research object and studies the comparative differences of consumer brand perception under the different sentiment and brand source classifications. The study was carried out with the aim to explore the influence mechanism of consumers' different brand attitudes. Structural Topic Modeling (STM) method was used to classify online reviews for brand perception topics, and Gephi network visualization was used to analyze the influence relationship between different brand perception topics. The study also conducts comparative research on the differences in perceived attitudes between positive and negative emotion classifications, as well as self-operated and non-self-operated brands, and analyzes the moderating effect of sentiment scores on the perceived theme intensity of different brands.