Multivariate redundancy analysis method for grouping after sparse matrix merging and implementation device of multivariate redundancy analysis method

The invention discloses a multivariate redundancy analysis method for grouping sparse matrixes after merging and an implementation device thereof, each DRAM (Dynamic Random Access Memory) storage area comprises a plurality of block groups, each block group comprises N blocks and has a shared column redundancy resource Ct, and each block has a row redundancy resource Rt; the method comprises the following steps of: merging fault points with the same row/column into a fault group FailGroup according to an incidence relation on the basis of fault points in each block to form M fault groups FailGroup; all elements of the fault group FailGroup 1 and the fault group FailGroup 2 are solved and merged to form a new fault group NewFailGroup, the solution space of the new fault group NewFailGroup is {R0Cx, R1Cx... RyCx... RtCx}, then the fault groups FailGroup 3 and NewFailGroup continue to be solved and merged until M fault groups are combined and completed, and the solution of Block is obtained; performing statistica