A Buffer Overflow Prediction Approach Based on Software Metrics and Machine Learning

Table 4

Mutual information value calculation results.

Metrics	C/C++ dataset		Java dataset
	Mutual information value	Sort	Mutual information value	Sort
CountInput	0.1892	11	0.6375	3
CountLine	0.5584	1	0.7996	1
CountLineCode	0.4902	2	0.6446	2
CountLineCodeDecl	0.4327	5	0.4373	11
CountLineCodeExe	0.3893	9	0.5	7
CountLineComment	0.3980	8	0.499	8
CountOutput	0.1625	12	0.3477	13
CountPath	Nan	13	0.3959	12
CountPathLog	Nan	13	0.2638	17
CountSemicolon	0.4307	7	0.5484	5
CountStmt	0.4465	4	0.5879	4
CountStmtDecl	0.4313	6	0.439	10
CountStmtExe	0.3434	10	0.5123	6
Cyclomatic	Nan	13	0.3122	14
CyclomaticModified	Nan	13	0.3122	15
CyclomaticStrict	Nan	13	0.2968	16
Essential	Nan	13	0.0099	20
Knots	Nan	13	0.2508	18
MaxEssentialKnots	Nan	13	0.0099	21
MaxNesting	Nan	13	0.2236	19
MinEssentialKnots	Nan	13	0.0099	22
RatioCommentToCode	0.4603	3	0.4959	9