Embodiments of the invention disclose an unbalanced data set processing method and device based on machine learning, relates to the technical field of data processing, and can solve the problem of distribution marginalization caused in the process of synthesizing an'artificial 'sample through an SMOTE algorithm. The unbalanced data set processing method based on machine learning comprises: obtaining a sample set according to a first sample set containing a plurality of types of image samples and an initial second sample set containing a plurality of types of image samples, generating a centralsample, the plurality of types of image samples and the few types of image samples including N-dimensional attributes, the central sample being composed of an average value of attributes of the plurality of types of image samples and the few types of image samples in each dimension, and N being greater than or equal to 1; and performing random linear interpolation between the central sample and the at least one minority class image sample to generate a newly-added minority class sample, and obtaining a second sample set updated by utilizing the newly-added minority class sample.