The invention relates to the technical field of food sensory evaluation, in particular to a food sensory experience evaluation and personalized recommendation
system based on electroencephalogram multi-
modal fusion, which comprises a food stimulation module, a
physiological reaction acquisition module, a computer
system, a
machine learning module and a personalized recommendation feedback module, according to the
system, multi-
modal physiological data such as electroencephalogram signals, electrocardiosignals and facial micro-expressions of a subject in the food stimulation process are collected, topological features are extracted by adopting a topological
data analysis method, emotion state manifold characterization is constructed, topology preserving mapping from the emotion state to food attributes is achieved, and the
food quality is improved. The core innovation of the invention lies in that a continuous coherence theory is adopted to process multi-
modal physiological signals, an 8-dimensional emotional state manifold is constructed, emotional-food attribute mapping is realized through a multilayer
radial basis function network, topology is designed to maintain constraints to ensure the stability of a structural relationship, and the system can objectively evaluate the sensory experience of food and improve the quality of food. The defect that a traditional evaluation method is high in subjectivity is overcome.