Knowledge discovery algorithm supporting vector prediction

The invention discloses a knowledge discovery algorithm for support vector prediction, which comprises the following steps: S1, acquiring an initial sample set, and initializing parameters; S2, training the sample set to obtain an RVM prediction model; S3, calculating a sample label, a local density factor and an error factor of each sample; S4, predicting a future sample to be input according tothe RVM prediction model; s5, arranging the sample characteristic vectors in a descending order and circulating, if the number of times of the non-correlation vectors exceeds a set threshold value, deleting the sample from the sample set, and jumping out of the circulation; s6, judging whether an input new sample exists or not, if the new sample exists, adding the new sample to form a new sample set, and going to the step S2; and if no new sample exists, outputting a predicted future sample. By introducing the concept of the sample label, the algorithm has higher prediction precision and shortvector prediction time, a