A Novel Graph Constructor for Semisupervised Discriminant Analysis: Combined Low-Rank and -Nearest Neighbor Graph

Table 4

Classification accuracy of different graphs with varying variance Gaussian noise.

Gaussian	0	0.02	0.04	0.06	0.08	0.1
LRKNN	0.901786	0.856786	0.845	0.851071	0.841071	0.826786
LR	0.620357	0.544643	0.543214	0.552857	0.543929	0.535357
SR-	0.811786	0.785714	0.799286	0.788214	0.780357	0.776786
SR	0.750714	0.558571	0.572857	0.588214	0.621429	0.629643
LLE-	0.751429	0.741071	0.724643	0.735	0.726429	0.727857
LLE	0.779286	0.725	0.713214	0.713214	0.717143	0.716429
KNNK	0.612487	0.547453	0.542545	0.548299	0.553657	0.556745
KNNB	0.645828	0.579001	0.571481	0.573906	0.563472	0.575386