Weight Design for Small Data Sample Decision under Attribute Association

In aircraft test, the sample size of flight test data is small, and the test data returned by telemetry also has the problem of attribute association. A weight design method of small sample decision under attribute association is proposed. The method enhances the small sample data based on the synthetic minority oversampling technique, reduces the dimension of variables through principal component analysis, normalizes the principal components based on the variance contribution rate, and obtains the weight of each principal component after dimensionality reduction. The research results can be applied to the weight design of the guidance and control system of aircraft, and provide a new method support for the subsequent relevant test evaluation.