The invention discloses an eSIM equipment
intelligent network selection method based on environment
perception and an AI strategy, and the method comprises the steps: collecting and preprocessing multi-source environment data, and generating a multi-mode original
feature vector; constructing a three-dimensional situation
tensor, generating a causal
structure diagram, clustering to obtain a situation identifier, and embedding a situation; constructing a situation feature table, improving TabNet to perform
feature selection, and outputting a high-dimensional situation representation vector; based on the situation characterization and the capability characteristics, performing
chain type prediction on the performance and generating an anti-factual income
score; constructing short-
time domain and long-
time domain preferences, generating a comprehensive preference value after fusion, and determining a target connection object; and an eSIM strategy configuration instruction is generated,
intelligent network selection switching is carried out, and
model parameters are incrementally updated. According to the method, active high-stability
intelligent network selection of the eSIM equipment in a complex scene is realized through fusion of multi-source environment
perception,
causal inference and an AI intelligent strategy.