27.06.2013 Views

6th European Conference - Academic Conferences

6th European Conference - Academic Conferences

6th European Conference - Academic Conferences

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Trojan.Sp<br />

y.Zeus.1.<br />

Gen.mal<br />

Trojan.Zb<br />

ot-<br />

1307.mal<br />

Trojan.Zb<br />

ot-<br />

2163.mal<br />

Tro<br />

jan.<br />

Sp<br />

y.Z<br />

eus<br />

.1.<br />

Ge<br />

n.<br />

mal<br />

54.<br />

23<br />

57.<br />

10<br />

63.<br />

53<br />

Troj<br />

an.Z<br />

bot-<br />

85.<br />

mal<br />

Madhu Shankarapani and Srinivas Mukkamala<br />

Troja<br />

n.Spy<br />

.Zeus<br />

.2.Ge<br />

n.mal<br />

63.8<br />

2 78.15<br />

58.4<br />

6 63.92<br />

56.5<br />

1 81.88<br />

Troj<br />

an.B<br />

roke<br />

r-<br />

12.<br />

mal<br />

61.3<br />

3<br />

78.4<br />

7<br />

84.0<br />

4<br />

5. Similarity analysis results<br />

Troj<br />

an.<br />

Zbo<br />

t-<br />

134<br />

2.m<br />

al<br />

86.<br />

42<br />

82.<br />

03<br />

74.<br />

65<br />

Troj<br />

an.<br />

Spy<br />

.Ze<br />

us.<br />

1.G<br />

en.<br />

mal<br />

66.<br />

27<br />

60.<br />

21<br />

55.<br />

75<br />

Troj<br />

an.S<br />

py.Z<br />

eus.<br />

1.Ge<br />

n.m<br />

al<br />

59.4<br />

5<br />

59.6<br />

4<br />

68.0<br />

1<br />

Troj<br />

an.S<br />

py.Z<br />

eus.<br />

1.Ge<br />

n.m<br />

al<br />

Troja<br />

n.Zbo<br />

t-<br />

290.<br />

mal<br />

Troja<br />

n.Spy<br />

.Zeus<br />

.1.Ge<br />

n.mal<br />

66.6<br />

9 76.85 58.62<br />

54.5<br />

9 64.83 64.50<br />

60.6<br />

6 50.08 60.47<br />

We apply the traditional similarity functions on Vs’ and Vu’. Cosine measure, extended Jaccard<br />

measure, and the Pearson correlation measure are the popular measures of similarity for sequences.<br />

The cosine measure is given below and captures a scale-invariant understanding of similarity.<br />

Cosine similarity: Cosine similarity is a measure of similarity between two vectors of n dimensions<br />

by finding the angle between them.<br />

Extended Jaccard measure: The extended Jaccard coefficient measures the degree of overlap<br />

between two sets and is computed as the ratio of the number of shared attributes ofVs’ AND Vu’to the<br />

number possessed byVs’ORVu’.<br />

Pearson correlation: Correlation gives the linear relationship between two variables. For a series of<br />

n measurements of variablesVs’andVu’, Pearson correlation is given by the formula below.<br />

Where and are values of variable Vs’ and Vu’ respectively at position i, n is the number of<br />

measurements, and are standard deviations of Vs’ and Vu’ respectively and and are<br />

means of Vs’ and Vu’ respectively.<br />

In these experiments, we calculated the mean value of the three measures. For a particular measure<br />

between a virus signature and a suspicious binary file, S(m)(Vs’i, Vu’), which stands for the similarity<br />

between virus signature i and a suspicious binary file. Our similarity report is generated by calculating<br />

the S(m)(Vs’i, Vu’) value for each virus signature in the signature database.<br />

In this experiment, we compared Zeus/Zbot variants against itself, creating n-by-n matrix which shows<br />

how similar are the variants. Table 1 shows the similarity values of Zeus/Zbot compared among<br />

themselves. From the Table 1 we can infer that variants of Zeus/Zbot are almost similar to sequence<br />

in which the Windows APIs are called.<br />

6. Conclusion<br />

In this paper, we present our effort of approach on malware detection based on Windows API call<br />

sequence. According to our observations, though there is tremendous increase in Zeus/Zbot variant<br />

builders, its behavior of API calls remains almost the same. Thus our approach can detect its variants<br />

258<br />

Troj<br />

an.Z<br />

bot-<br />

115<br />

1.m<br />

al<br />

74.0<br />

3<br />

73.5<br />

8<br />

78.3<br />

0<br />

DHL<br />

_DO<br />

C.m<br />

al<br />

(3)<br />

80.9<br />

5<br />

64.2<br />

9<br />

50.7<br />

3<br />

(1)<br />

(2)<br />

Troj<br />

an.<br />

Spy.<br />

Zeu<br />

s.1.<br />

Gen<br />

.mal<br />

87.5<br />

1<br />

54.3<br />

4<br />

56.1<br />

0

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!