The invention relates to the technical field of computers, in particular to a
test case generalization method for automobile network and
data security, which comprises the following steps: collecting and preprocessing multi-source heterogeneous data, fusing the data to construct a dynamic risk
feature model to extract an
attack vector, and constructing a layered scene space based on real-time topology of a vehicle-mounted network; a two-channel generalization engine is utilized, unknown
threat scene cases are generated by means of a structure
causal model, and variation is injected into nodes of a protocol
syntax tree to generate fuzzy test cases so as to enrich case types; a
reinforcement learning framework is introduced to dynamically optimize a
test path, and a closed-loop feedback mechanism is adopted to expand high-risk scene nodes, so that the multi-source heterogeneous
data processing capability is effectively improved, the unknown
threat simulation capability is enhanced, the
test efficiency is improved, and the problem that a test method in the prior art is incomplete in multi-source heterogeneous
data processing and poor in reliability is solved. The technical problems of influence on subsequent analysis accuracy,
single test case generation type, lack of unknown
threat simulation and low
test efficiency in the prior art are solved.