Algorithm 2 Drug resistance classification algorithm
1:	repeat
2:	Randomly choose m vectors from SQV group, the rest n being training data
3:	Construct dictionary J using Algorithm 1
4:	Randomly choose m vectors from none group, the rest n being training data
5:	Construct dictionary K using Algorithm 1
6:	for each vector v in testing data do
7:	computing the sparse representation of v using both dictionaries J and K
8:	computing the representation errors using the two dictionaries
9:	if the error of using J is larger then
10:	v is resistant to SQV
11:	else
12:	v is NOT resistant to SQV
13:	end if
14:	end for
15:	Compute the confusion matrix
16:	until For 9 times