This is a Data Interpretation Document.
1 ,mnist_test_set
mnist_data.mat : 784*10000
mnist_data_1.mat :784*10000
2 ,qmnist_test_set
Qmnist_data.mat :784*60000
Qmnist_data_1.mat :784*60000
MNIST / (MNIST _1) QMNIST /(QMNIST _1) Real label
Col-mark ,num ,Col-mark ,num
1-980 ,980 ,1-5952 ,5952 ,0
981-2115 ,1135 ,5953-12743, 6791 ,1
2116-3147 ,1032 ,12744-18769 ,6026 ,2
3148-4157 ,1010 ,18770-24853 ,6084 ,3
4158-5139 ,982 ,24854-30633 ,5780 ,4
5140-6031 ,892 ,30634-36087, 5454 ,5
6032-6989 ,958 ,36088-42044 ,5957 ,6
6990-8017 ,1028 ,42045-48275, 6231 ,7
8018-8991 ,974, 48276-54165 ,5890 ,8
8992-10000 ,1009, 54166-60000 ,5835 ,9
Total ,10000 ,60000
3 ,00 ,01 ,02 ,03 ,04 are the sample data generated based on SPCA.
0.mat : training the data “0” which is the 1st column of samp_data.mat
lossall: 2*5 cell——2 layers and 5 patterns——loss value
recognoresult : 5*2 cell
1st column ——Unrecognized labels at each layer in mnist_data.
2nd column ——Unrecognized labels at each layer in mnist_data_1.
recognoresultq: 5*2 cell
1st column ——Unrecognized labels at each layer in qmnist_data.
2nd column ——Unrecognized labels at each layer in qmnist_data_1.
recogresult: 5*5 cell ,row represents pattern
1st column ——recognized labels at each layer in mnist_data.
2nd column ——Recognition Rate of 0-9 in mnist_data.
3rd & 4th col——Total number of identifications per layer.
5th col ——Number of correct identifications for each number.
recogresult_1: For mnist_data_1.
recogresult_1q: For qmnist_data_1.
recogresultq : For qmnist_data.
res_imgdata: 2*5 cell——2 layers and 5 patterns——impression
timedata: Reference time (s) for block operation.
4 ,samp_data_origin1 —— samp_data_origin10
Digital data sampled from the original training set（mnist_train_set,60000）
5 test_results
1),Validation_00_mnist_data.mat 10000*3
training “0” in files <00>, test on mnist_data.
1st col : real label
2nd col : Recognition results of 1st layer under P3.
3rd col： Recognition results of 2nd layer under P3.
2), Validation_00_mnist_data_1 ,For mnist_data_1
3), Validation_00_Qmnist_data ,For Qmnist_data
4), Validation_00_Qmnist_data_1 ,For Qmnist_data_1
6, \data\00\impressiondata\
impression_data.mat , 2*10 cell ,the results of intermediate process.