Download MULTICLASS SUPPORT VECTOR MACHINES: A COMPARATIVE …

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Document related concepts

Nonlinear dimensionality reduction wikipedia, lookup

K-nearest neighbors algorithm wikipedia, lookup

Cluster analysis wikipedia, lookup

Transcript
Data Mining Techniques for
Malware Detection
R. K. Agrawal
School of Computer and Systems Sciences
Jawaharlal Nehru University
NewDelhi-110067
1
Outline
• Data Mining
• Classification
• Clustering
• Association Rules
• Experimental Results
• Conclusion and Future Work
2
Motivation:
“Necessity is the Mother of Invention
• Data explosion problem
– Automated data collection tools lead to tremendous amounts of data
stored in databases and other information repositories
• We are drowning in data, but starving for knowledge!
• Solution: data mining
– Extraction of interesting knowledge (rules, regularities, patterns,
constraints) from data in large databases
3
Commercial Viewpoint
• Lots of data is being collected
and warehoused
– Web data, e-commerce
– purchases at department/
grocery stores
– Bank/Credit Card
transactions
• Computers have become cheaper and more powerful
• Competitive Pressure is Strong
– Provide better, customized services for an edge (e.g. in
Customer Relationship Management)
4
Scientific Viewpoint
• Data collected and stored at
enormous speeds (GB/hour)
– remote sensors on a satellite
– Network related Log files
– microarrays generating gene
expression data
– scientific simulations
generating terabytes of data
• Traditional techniques infeasible for raw
data
• Data mining may help scientists
– in classifying and segmenting data
– in Hypothesis Formation
5
What Is Data Mining?
• Data mining (knowledge discovery in databases):
– Extraction of interesting (non-trivial, implicit, previously
unknown and potentially useful) information or patterns
from data in large databases
• Alternative names:
– Knowledge discovery(mining) in databases (KDD),
knowledge extraction, data/pattern analysis, data
archeology, business intelligence, etc.
6
Data Mining Tasks
• Prediction Tasks
– Use some variables to predict unknown or future values of other
variables
• Description Tasks
– Find human-interpretable patterns that describe the data.
Common data mining tasks
–
–
–
–
–
–
Classification [Predictive]
Clustering [Descriptive]
Association Rule Discovery [Descriptive]
Sequential Pattern Discovery [Descriptive]
Regression [Predictive]
Deviation Detection [Predictive]
7
Classification: Definition
• Given a collection of records (training set )
– Each record contains a set of attributes, one of the attributes is
the class label.
• Find a model for class attribute as a function of
the values of other attributes. q : X  Y
• Goal: previously unseen records should be
assigned a class as accurately as possible.
8
Classification—A Two-Step Process
• Model construction: describing a set of predetermined classes
– Each tuple/sample is assumed to belong to a predefined class,
as determined by the class label attribute
– The set of tuples used for model construction is training set
– The model is represented as classification rules, decision trees,
or mathematical formulae
• Model usage: for classifying future or unknown objects
– Estimate accuracy of the model
• The known label of test sample is compared with the
classified result from the model
• Accuracy rate is the percentage of test set samples that are
correctly classified by the model
– If the accuracy is acceptable, use the model to classify data
tuples whose class labels are not known
9
Process (1): Model Construction
Training
Data
NAME
Rahul
Mohan
Dev
Kirti
Sudhir
Arun
RANK
YEARS TENURED
Assistant Prof
3
no
Assistant Prof
7
yes
Professor
2
yes
Associate Prof
7
yes
Assistant Prof
6
no
Associate Prof
3
no
Classification
Algorithms
Classifier
(Model)
IF rank = ‘professor’
OR years > 6
THEN tenured = ‘yes’
10
Process (2): Using the Model in Prediction
Classifier
Testing
Data
Unseen Data
(Jolly Professor, 5)
NAME
Tarun
Manish
Jolly
Harsh
RANK
YEARS TENURED
Assistant Prof
2
no
Associate Prof
7
no
Professor
5
yes
Assistant Prof
7
yes
Tenured?
11
Classification: Application
• Malware Detection
– Goal: Predict whether the given binary is Malware or
not.
– Approach:
• Use both kind of binaries (Normal and Malware)
• Learn a model for the class of the binaries.
• Use this model to detect malware by observing a binary.
12
Clustering Definition
• Given a set of data points, each having a set of
attributes, and a similarity measure among
them, find clusters such that
– Data points in one cluster are more similar to one
another.
– Data points in separate clusters are less similar to one
another.
• Similarity Measures:
– Euclidean Distance if attributes are continuous.
– Other Problem-specific Measures.
13
Illustrating Clustering
x Euclidean Distance Based Clustering in 3-D space.
Intracluster distances
are minimized
Intercluster distances
are maximized
14
Clustering: Application
• Binaries Segmentation:
– Goal: subdivide a given set of binaries into distinct
subsets of binaries
15
Association Rule Discovery: Definition
• Given a set of records each of which contain some
number of items from a given collection;
– Produce dependency rules which will predict occurrence of an
item based on occurrences of other items.
TID
Items
1
2
3
4
5
Bread, Coke, Milk
Beer, Bread
Beer, Coke, Diaper, Milk
Beer, Bread, Diaper, Milk
Coke, Diaper, Milk
Rules Discovered:
{Bread} --> {Milk}
{Diaper} --> {Beer}
16
The Sad Truth About Diapers and Beer
• So, don’t be surprised if you find six-packs stacked next to diapers!
17
Association Rule Discovery: Application
• Malware Rules
– Goal: To identify activities that are happen together in
a given malware.
18
Sequential Pattern Discovery:
Definition
Given is a set of objects, with each object associated with
its own timeline of events, find rules that predict strong
sequential dependencies among different events:
– In telecommunications alarm logs,
• (Inverter_Problem Excessive_Line_Current)
(Rectifier_Alarm) --> (Fire_Alarm)
– In point-of-sale transaction sequences,
• Computer Bookstore:
(Intro_To_Visual_C) (C++_Primer) -->
(Perl_for_dummies)
• Athletic Apparel Store:
(Shoes) (Racket, Racketball) --> (Sports_Jacket)
19
Classification Example
height
weight
 x1 
x   x
 2
X  2
Training examples {( x1 , y1 ), , (xl , yl )}
Linear classifier:
H if (w  x)  b  0
q(x)  
 J if (w  x)  b  0
x2
yH
y  J ( w  x)  b  0
20
x1
Classification Techniques
•
•
•
•
•
•
Decision Trees
Naïve Bayes
Support Vector Machines
Neural Networks
Parzen Window
K-nearest neigbor
21
Issues: Data Preparation
• Data cleaning
– Preprocess data in order to reduce noise and handle
missing values
• Relevance analysis (feature selection)
– Remove the irrelevant or redundant attributes
• Data transformation
– Generalize and/or normalize data
22
Issues: Evaluating Classification Methods
• Accuracy
– classifier accuracy: predicting class label
– predictor accuracy: guessing value of predicted
attributes
• Speed
– time to construct the model (training time)
– time to use the model (classification/prediction time)
• Robustness: handling noise and missing values
• Scalability: efficiency in disk-resident databases
• Interpretability
– understanding and insight provided by the model
• Other measures, e.g., goodness of rules, such as decision
tree size or compactness of classification rules
23
Decision Tree Induction: Training Dataset
This
follows an
example
of
Quinlan’s
ID3
age
<=30
<=30
31…40
>40
>40
>40
31…40
<=30
<=30
>40
<=30
31…40
31…40
>40
income student credit_rating
high
no fair
high
no excellent
high
no fair
medium
no fair
low
yes fair
low
yes excellent
low
yes excellent
medium
no fair
low
yes fair
medium
yes fair
medium
yes excellent
medium
no excellent
high
yes fair
medium
no excellent
buys_computer
no
no
yes
yes
yes
no
yes
no
yes
yes
yes
yes
yes
no
24
A Decision Tree for “buys_computer”
age?
<=30
31..40
overcast
student?
no
no
yes
yes
yes
>40
credit rating?
excellent
fair
yes
25
Algorithm for Decision Tree Induction
• Basic algorithm (a greedy algorithm)
– Tree is constructed in a top-down recursive divide-and-conquer
manner
– At start, all the training examples are at the root
– Attributes are categorical (if continuous-valued, they are
discretized in advance)
– Examples are partitioned recursively based on selected attributes
– Test attributes are selected on the basis of a heuristic or
statistical measure (e.g., information gain)
• Conditions for stopping partitioning
– All samples for a given node belong to the same class
– There are no remaining attributes for further partitioning –
majority voting is employed for classifying the leaf
– There are no samples left
26
Attribute Selection Measure:
Information Gain (ID3/C4.5)



Select the attribute with the highest information gain
Let pi be the probability that an arbitrary tuple in D
belongs to class Ci, estimated by |Ci, D|/|D|
Expected information (entropy) needed to classify a tuple
m
in D:
Info( D)   pi log 2 ( pi )
i 1


Information needed (after using A to split D into v
v |D |
partitions) to classify D:
j
InfoA ( D)  
 I (D j )
j 1 | D |
Information gained by branching on attribute A
Gain(A)  Info(D)  InfoA(D)
27
Attribute Selection: Information Gain


Class P: buys_computer = “yes”
Class N: buys_computer = “no”
Info( D)  I (9,5)  
age
<=30
31…40
>40
age
<=30
<=30
31…40
>40
>40
>40
31…40
<=30
<=30
>40
<=30
31…40
31…40
>40
Infoage ( D) 
9
9
5
5
log 2 ( )  log 2 ( ) 0.940
14
14 14
14
pi
2
4
3
ni I(pi, ni)
3 0.971
0 0
2 0.971
income student credit_rating
high
no
fair
high
no
excellent
high
no
fair
medium
no
fair
low
yes fair
low
yes excellent
low
yes excellent
medium
no
fair
low
yes fair
medium
yes fair
medium
yes excellent
medium
no
excellent
high
yes fair
medium
no
excellent
buys_computer
no
no
yes
yes
yes
no
yes
no
yes
yes
yes
yes
yes
no

5
4
I (2,3) 
I (4,0)
14
14
5
I (3,2)  0.694
14
5
I (2,3) means “age <=30” has 5
14
out of 14 samples, with 2 yes’es
and 3 no’s. Hence
Gain(age)  Info( D)  Infoage ( D)  0.246
Similarly,
Gain(income)  0.029
Gain( student )  0.151
Gain(credit _ rating )  0.048
28
Computing Information-Gain for
Continuous-Value Attributes
• Let attribute A be a continuous-valued attribute
• Must determine the best split point for A
– Sort the value A in increasing order
– Typically, the midpoint between each pair of adjacent
values is considered as a possible split point
• (ai+ai+1)/2 is the midpoint between the values of ai and ai+1
– The point with the minimum expected information
requirement for A is selected as the split-point for A
• Split:
– D1 is the set of tuples in D satisfying A ≤ split-point, and
D2 is the set of tuples in D satisfying A > split-point 29
Linear Classifiers
denotes +1
f(x,w,b) = sign(w x + b)
denotes -1
Any of these
would be fine..
..but which is
best?
30
Support Vector Machine
x+
M=Margin Width
Support Vectors
are those
datapoints that
the margin
pushes up
against
X-
What we know:
• w . x+ + b = +1
• w . x- + b = -1
• w . (x+-x-) = 2


(x  x )  w 2
M 

w
w
31
Linear SVM Mathematically
•
Goal: 1) Correctly classify all training data
wxi  b  1
wxi  b  1
yi (wxi  b)  1
2) Maximize
the Margin M
2

w
same as minimize
•
if yi = +1
If yi = -1
for all i
1 t
ww
2
We can formulate a Quadratic Optimization Problem and solve for w and b
Minimize
subject to
1 t
( w)  w w
2
yi (wxi b)  1 i
32
Linear SVM. Cont.
•
Requiring the derivatives with respect to w,b to vanish yields:
m
maximize

1 m m
i   j  i   j 




y
y x ,x


i
i j
2 i 1 j 1
i 1
m
Subject to :
 y   0
i
i 1
i
 i  0 i
•
KKT conditions yield:
•
Where:
for any  i  0, b  y i   w, x i 
m
w    i y i  x i 
i 1
33
Linear SVM. Cont.
• The resulting separating function is:
m
sgn f  x   sgn   i y i  x i  , x  b
i 1
34
Linear SVM. Cont.
•
Requiring the derivatives with respect to w,b to vanish yields:
m
maximize

1 m m
i   j  i   j 




y
y x ,x


i
i j
2 i 1 j 1
i 1
m
Subject to :
 y   0
i
i 1
i
 i  0 i
•
KKT conditions yield:
•
Where:
for any  i  0, b  y i   w, x i 
m
w    i y i  x i 
i 1
35
Linear SVM. Cont.
• The resulting separating function is:
m
sgn f  x   sgn   i y i  x i  , x  b
i 1
• Notes:
– The points with α=0 do not affect the solution.
– The points with α≠0 are called support vectors.
– The equality conditions hold true only for the Support Vectors.
36
Non-separable case
•
The modifications yield the following problem:
m
maximize

1 m m
i   j  i   j 




y
y x ,x


i
i j
2 i 1 j 1
i 1
m
Subject to :
i 

y
 i 0
i 1
0  i  C
i
37
Non Linear SVM
• Note that the training data appears in the solution only in inner
products.
• If we pre-map the data into a higher and sparser space we can get
more separability and a stronger separation family of functions.
• The pre-mapping might make the problem infeasible.
• We want to avoid pre-mapping and still have the same separation
ability.
• Suppose we have a simple function that operates on two training
points and implements an inner product of their pre-mappings, then
we achieve better separation with no added cost.
38
Non-linear SVMs: Feature spaces
• General idea: the original feature space can always be mapped to
some higher-dimensional feature space where the training set is
separable:
Φ: x → φ(x)
39
The “Kernel Trick”
•
•
•
•
•
The linear classifier relies on inner product between vectors K(xi,xj)=xiTxj
If every datapoint is mapped into high-dimensional space via some
transformation Φ: x → φ(x), the inner product becomes:
K(xi,xj)= φ(xi) Tφ(xj)
A kernel function is a function that is equivalent to an inner product in some
feature space.
Example:
2-dimensional vectors x=[x1 x2]; let K(xi,xj)=(1 + xiTxj)2,
Need to show that K(xi,xj)= φ(xi) Tφ(xj):
K(xi,xj)=(1 + xiTxj)2,= 1+ xi12xj12 + 2 xi1xj1 xi2xj2+ xi22xj22 + 2xi1xj1 + 2xi2xj2=
= [1 xi12 √2 xi1xi2 xi22 √2xi1 √2xi2]T [1 xj12 √2 xj1xj2 xj22 √2xj1 √2xj2]
=
= φ(xi) Tφ(xj), where φ(x) = [1 x12 √2 x1x2 x22 √2x1 √2x2]
Thus, a kernel function implicitly maps data to a high-dimensional space
(without the need to compute each φ(x) explicitly).
40
Mercer Kernels
•
•
A Mercer kernel is a function:
k:Xd Xd R
for which there exists a function:
: Xd  H
such that:
k ( x, y)  ( x), ( y)
x, y  X d
A function k(.,.) is a Mercer kernel if
for any function g(.), such that:
the following holds true:
2
g
 ( x)dx  
  g ( x) g ( y)k ( x, y)dxdy  0
41
Commonly used Mercer Kernels
•
Homogeneous Polynomial Kernels:
•
Non-homogeneous Polynomial Kernels:
•
Radial Basis Function (RBF) Kernels:
k ( x, y )   x, y

p
k ( x, y )   x, y  1

p
k ( x, y )  exp   x  y
2

42
Solution of non-linear SVM
•
The problem:

m
1 m m
i   j 
i   j 




y
y
k
x
,x


i
i j
2 i 1 j 1
i 1
maximize


m
Subject to :
i 

y
 i 0
i 1
•
The separating function:
0  i  C
i
m


sgn f  x   sgn   i y i k x i  , x  b
i 1
43
Multi-Class SVM
• Approaches:
• One against One ( K (K-1) / 2 ) binary Classifiers required
Outputs of the classifiers are aggregated to make the final decision.
• One against All (K binary Classifiers required):
It trains k binary classifiers, each of which separates one class from
the other (k-1) classes. Given a data point X , the binary classifier
with the largest output determines the class of X.
44
Why Is SVM Effective on High Dimensional Data?

The complexity of trained classifier is characterized by the # of
support vectors rather than the dimensionality of the data

The support vectors are the essential or critical training examples —
they lie closest to the decision boundary (MMH)

If all other training examples are removed and the training is
repeated, the same separating hyperplane would be found

The number of support vectors found can be used to compute an
(upper) bound on the expected error rate of the SVM classifier, which
is independent of the data dimensionality

Thus, an SVM with a small number of support vectors can have good
generalization, even when the dimensionality of the data is high
45
Experiments
• Source of data: Preprocessed data in terms of API
Calls taken from data collected from C-Dac Mohali.
• Description of data
Sample
Space
Training
set
Testing
set
Benign
534
50
484
Malicious
168
50
118
Total
702
100
602
46
Classifier Accuracy Measures
C1
C2
C1
True positive
False negative
C2
False positive
True negative
• Performance measures
sensitivity = t-pos/pos
specificity = t-neg/neg
/* true positive recognition rate */
/* true negative recognition rate */
accuracy = sensitivity * pos/(pos + neg) + specificity * neg/(pos + neg)
47
Experimental Results
Classifier
sensitivity
specificity
k=5
K=6
K=7
K=5
K=6
K=7
C4.5
70.86
71.23
69.68
68.62
69.96
61.05
SVM
75.26
76.79
75.18
73.54
78.34
74.46
48
Observations
•
The performance of SVM classifier is significantly
better in comparison to C4.5.
• The performance is dependent on the size of
feature size
• SVM requires less training samples in comparison
C4.5. Hence, svm is a better choice as collecting
malicious samples is difficult.
49
Conclusion & Future Work
• SVM is a better classification technique which can
be used for detection of Malware.
• Needs attention to construct better feature
representation for better generalization
• How to extend it to multi-class malware problem
50
References
•
C. J. C. Burges. A Tutorial on Support Vector Machines for Pattern Recognition.
Data Mining and Knowledge Discovery, 2(2): 121-168, 1998.
•
J. R. Quinlan. C4.5: Programs for Machine Learning. Morgan Kaufmann, 1993.
•
P. Tan, M. Steinbach, and V. Kumar. Introduction to Data Mining. Addison Wesley,
2005.
•
I. H. Witten and E. Frank. Data Mining: Practical Machine Learning Tools and
Techniques, 2ed. Morgan Kaufmann, 2005.
•
Han and Kamber, Data Mining Concepts
•
B. Zhang, J. Yin, J. Hao, D. Zhang, S. Wang, Using Support Vector Machine to detect
unknown computer viruses, Int. Journal of Computational Intelligence Research, vol. 2,
No. 1, pp. 100-104, 2006.
Szappanos,G.: Are There Any Polymorphic Macro Viruses at ALL (and What to Do with
Them).in Proceedings of the 12th International Virus Bulletin Conference, 2001.
Forrest,S., Hofmeyr, S. A., Somayaji, A.: Computer immunology. Communications of the
ACM. 10, pp. 88–96, 1997.
Lee,W., Dong,X.: Information-Theoretic measures for anomaly detection. In: Needham,R.,
Abadi M, (eds):. Proceedings of the 2001 IEEE Symposium on Security and Privacy
Oakland, CA: IEEE Computer Society Press, pp. 130-143, 2001.
51
LIBSVM. http://www.csie.ntu.edu.tw/~cjlin/.
•
•
•
•
References (4)
•
P. Tan, M. Steinbach, and V. Kumar. Introduction to Data Mining. Addison Wesley,
2005.
•
S. M. Weiss and C. A. Kulikowski. Computer Systems that Learn: Classification
and Prediction Methods from Statistics, Neural Nets, Machine Learning, and
Expert Systems. Morgan Kaufman, 1991.
•
S. M. Weiss and N. Indurkhya. Predictive Data Mining. Morgan Kaufmann, 1997.
•
I. H. Witten and E. Frank. Data Mining: Practical Machine Learning Tools and
Techniques, 2ed. Morgan Kaufmann, 2005.
•
X. Yin and J. Han. CPAR: Classification based on predictive association rules.
SDM'03
•
H. Yu, J. Yang, and J. Han. Classifying large data sets using SVM with
hierarchical clusters. KDD'03.
52