Realization method of lightweight super-dimensional calculation classification model

The invention discloses an implementation method of a lightweight super-dimensional calculation classification model, which comprises the following steps of: firstly, determining the data volume, the data value domain and the super-dimensional vector dimension of a sample to be learned, dividing an address memory and a continuous item memory in a hardware memory, determining the number of numerical sampling, and obtaining an address super-dimensional vector; obtaining a numerical value super-dimensional vector through a NAND gate negation method; for an input sample, firstly, an address super-dimensional vector of the position where each data point is located and a numerical value super-dimensional vector corresponding to the numerical value of the data point are bound through multiplication, and then all the data points form a set through additive operation and are subjected to binarization processing; when the sample is a training sample, the sample is added to the category super-dimensional vector correspo