Urdu Handwritten Characters Data Visualization and Recognition Using Distributed Stochastic Neighborhood Embedding and Deep Network

Table 2

The confusion matrix showing accuracy rate of classification of Urdu handwritten numeral using 8-fold cross validation.

Eight	Five	Four	Nine	One	Seven	Six	Three	Two	Zero	Classified
8	5	4	9	1	7	6	3	2	0	As
94%	3%	0	0	0	0	0	0	0	3%	Eight
3%	95%	0	0	0	0	0	0	0	2%	Five
0	0	94%	0	0	0	2%	3%	1%	0	Four
0	0	0	95%	3%	0	2%	0	0	0	Nine
0	0	0	3%	96%	0	1%	0	0	0	One
0	4%	0	0	0	96%	0	0	0	0	Seven
0	0	0	2	4%	0	93%	0	0	3%	Six
0	0	2%	0	0	0	0	96%	2%	0	Three
0	0	3%	0	4%	0	0	1%	92%	0	Two
1%	3%	0	0	0	0	0	0	0	96%	Zero

The values in bold (in diagonal) are representing the accuracy percentage of the classification of Urdu handwritten numerals obtained by applying using variations of n-fold cross-validation approach.