Detection of false data injection attacks on power systems using graph edge-conditioned convolutional networks

Protection and Control of Modern Power Systems

Table 3 Detection Results of Different Attack Intensities and Attack Degrees

Attack	Low attack intensity				Medium attack intensity				High attack intensity
Degrees	Acc	Pre	Reca	F₁	Acc	Pre	Reca	F₁	Acc	Pre	Reca	F₁
\(\gamma _1\)	0.9918	0.9955	0.9109	0.9513	0.9941	0.9938	0.9418	0.9671	0.9940	0.9926	0.9408	0.9660
\(\gamma _2\)	0.9976	0.9916	0.9813	0.9864	0.9995	0.9939	1.0000	0.9969	0.9995	0.9951	0.9990	0.9971
\(\gamma _3\)	0.9992	0.9931	0.9980	0.9955	0.9995	0.9939	1.0000	0.9970	0.9998	0.9980	1.0000	0.9990
\(\gamma _4\)	0.9989	0.9886	0.9990	0.9938	0.9990	0.9888	1.0000	0.9944	0.9994	0.9927	1.0000	0.9964
\(\gamma _5\)	0.9993	0.9941	0.9980	0.9961	0.9993	0.9921	1.0000	0.9960	0.9994	0.9930	1.0000	0.9965
\(\gamma _6\)	0.9993	0.9939	0.9980	0.9960	0.9995	0.9940	1.0000	0.9970	0.9993	0.9918	1.0000	0.9959
\(\gamma _7\)	0.9992	0.9924	0.9990	0.9957	0.9998	0.9980	1.0000	0.9990	0.9992	0.9913	1.0000	0.9956
\(\gamma _8\)	0.9993	0.9922	1.0000	0.9961	0.9997	0.9971	1.0000	0.9985	0.9994	0.9928	1.0000	0.9964
\(\gamma _9\)	0.9993	0.9942	0.9981	0.9961	0.9995	0.9937	1.0000	0.9968	0.9995	0.9941	1.0000	0.9971
\(\gamma _{10}\)	0.9993	0.9930	0.9990	0.9960	0.9996	0.9961	1.0000	0.9980	0.9998	0.9980	1.0000	0.9990
\(\gamma _{11}\)	0.9993	0.9916	1.0000	0.9958	0.9995	0.9941	1.0000	0.9970	0.9993	0.9919	1.0000	0.9959
\(\gamma _{12}\)	0.9984	0.9937	0.9875	0.9906	0.9997	0.9969	1.0000	0.9985	0.9993	0.9923	1.0000	0.9961
\(\gamma _{13}\)	0.9982	0.9927	0.9866	0.9896	0.9995	0.9937	1.0000	0.9968	0.9992	0.9912	1.0000	0.9956
\(\gamma _{14}\)	0.9979	0.9937	0.9824	0.9880	0.9995	0.9949	1.0000	0.9974	0.9994	0.9930	1.0000	0.9965