Traditional brand design methods are difficult to accurately grasp audience needs, and data mining technology provides a new way to solve this problem. This paper constructs a brand design demand model for financial central enterprises based on data mining methods. The study adopts web crawler to obtain brand review text, applies word2vec to realize word vectorization; uses TF-IDF algorithm to extract brand overall imagery; analyzes brand local imagery based on syntactic relationship; realizes perceptual imagery parameterization through word vector technology; establishes brand demand element model, analyzes the correlation between brand features, emotional features, contextual features and behavioral features. The study found that “national credit” has the highest word frequency (14520) and TF-IDF value of 0.02687 among the theme words processed by data mining; the social factor has the highest weight (0.2278) in the brand design evaluation system constructed; and the W brand of a financial central enterprise designed by applying the model obtained a score of 4.39, with the highest score (4.44) for the spiritual factor. The study establishes a brand design model driven by “visual” + “semantic” dual modes, which improves the relevance and accuracy of the brand design of financial centralized enterprises, and provides a scientific method and practical path for the brand design of financial centralized enterprises.
