Cascade support vector transfer learning method based on Gaussian mixture weighting

The invention discloses a cascaded support vector transfer learning method based on Gaussian mixture weighting. The method comprises the following steps: S1, initializing a sample weight; s2, constructing a weak classifier set based on the classification hyperplane; s3, calculating the overall error rate of the weak classifier; s4, constructing a weight regulation factor of the current iteration step based on the error rate; s5, constructing a case weight vector; s6, a Gaussian mixture-based weight weighting strategy and a weight two-branch transmission strategy; and S7, gradually completing transfer learning of the whole model by using an iteration strategy. According to the method, a more accurate model can be obtained only by using a small amount of new data and a large amount of old data, overfitting caused by too fast case learning is avoided by using a Gaussian mixture weighting strategy, meanwhile, forward learning in a migration process is ensured by using a residual network structure, the problem of