10/24/08 1 Anomaly Detection Using Anomaly Detection Using Data Mining Techniques Data Mining Techniques Margaret H. Dunham, Yu Meng, Margaret H. Dunham, Yu Meng, Donya Quick, Jie Huang, Charlie Donya Quick, Jie Huang, Charlie Isaksson Isaksson CSE Department CSE Department Southern Methodist University Southern Methodist University Dallas, Texas 75275 Dallas, Texas 75275 [email protected]This material is based upon work supported by the National This material is based upon work supported by the National Science Foundation under Grant No. Science Foundation under Grant No. IIS-0208741 IIS-0208741
44
Embed
Anomaly Detection Using Data Mining Techniques, SMU CSE ...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
10/24/08 1
Anomaly Detection Using Data Anomaly Detection Using Data Mining TechniquesMining Techniques
Margaret H. Dunham, Yu Meng, Donya Margaret H. Dunham, Yu Meng, Donya Quick, Jie Huang, Charlie IsakssonQuick, Jie Huang, Charlie Isaksson
CSE DepartmentCSE DepartmentSouthern Methodist UniversitySouthern Methodist University
This material is based upon work supported by the National Science This material is based upon work supported by the National Science Foundation under Grant No. Foundation under Grant No. IIS-0208741 IIS-0208741
Develop modeling techniques which can “learn/forget” past behavior of spatiotemporal stream events. Apply to prediction of anomalous events.
Introduction EMM Overview EMM Applications to Anomaly Detection Future Work
10/24/08 3
10/24/08 4
Outline
Introduction Motivation What is an anomaly? Spatiotemporal Data Modeling Spatiotemporal Data
EMM Overview EMM Applications to Anomaly Detection Future Work
10/24/08 5
MotivationA growing number of applications generate streams of data.
Computer network monitoring data Call detail records in telecommunications Highway transportation traffic data Online web purchase log records Sensor network data Stock exchange, transactions in retail chains, ATM operations in
banks, credit card transactions.Data mining techniques play a key role in modeling and
analyzing this data.
10/24/08 6
What is Anomaly?
Event that is unusual Event that doesn’t occur frequently Predefined event What is unusual? What is deviation?
10/24/08 7
What is Anomaly in Stream Data?
Rare - Anomalous – Surprising Out of the ordinary Not outlier detection
No knowledge of data distribution Data is not static Must take temporal and spatial values into account May be interested in sequence of events
Ex: Snow in upstate New York is not an anomaly Snow in upstate New York in June is rare
Rare events may change over time
10/24/08 8
Statistical View of Anomaly Outlier Data item that is outside the normal distribution of the
data Identify by Box Plot
Image from Image from Data Mining, Introductory and Advanced TopicsData Mining, Introductory and Advanced Topics, Prentice Hall, 2002., Prentice Hall, 2002.
10/24/08 9
Statistical View of Anomaly
Image from Image from www.wikipedia.org, Normal distribution..
Data Mining View of Anomaly Classification Problem
Build classifier from training data Problem is that training data shows what is NOT an
anomaly Thus an anomaly is anything that is not viewed as
normal by the classification technique MUST build dynamic classifier
Identify anomalous behavior Signatures of what anomalous behavior looks like Input data is identified as anomaly if it is similar enough
to one of these signatures Mixed – Classification and Signature
10/24/08 11
Visualizing Anomalies Temporal Heat Map (THM) is a visualization technique for streaming
data derived from multiple sensors. Two dimensional structure similar to an infiniteinfinite table. Each row of the table is associated with one sensor value. Each column of the table is associated with a point in time. Each cell within the THM is a color representation of the sensor value Colors normalized (in our examples)
0 – While 0.5 – Blue 1.0 - Red
10/24/08 12
THM of VoIP Data
VoIP traffic data was provided by Cisco Systems and represents logged VoIP traffic in their Richardson, Texas facility from Mon Sep 22 12:17:32 2003 to Mon Nov 17 11:29:11 2003.
10/24/08 13
Spatiotemporal Stream DataRecords may arrive at a rapid rateHigh volume (possibly infinite) of continuous dataConcept drifts: Data distribution changes on the flyData does not necessarily fit any distribution patternMultidimensionalTemporalSpatialData are collected in discrete time intervals,Data are in structured format, <a1, a2, …>Data hold an approximation of the Markov property.
10/24/08 14
Spatiotemporal Environment Events arriving in a stream At any time, t, we can view the state
of the problem as represented by a vector of n numeric values:
Vt = <S1t, S2t, ..., Snt>
V1 V2 … VqS1 S11 S12 … S1q
S2 S21 S22 … S2q
… … … … …Sn Sn1 Sn2 … SnqTime
10/24/08 15
Data Stream Modeling
Single pass: Each record is examined at most once Bounded storage: Limited Memory for storing synopsis Real-time: Per record processing time must be low Summarization (Synopsis )of data Use data NOT SAMPLE Temporal and Spatial Dynamic Continuous (infinite stream) Learn Forget Sublinear growth rate - Clustering
11/26/07 – IRADSN’07 15
10/24/08 16
MMA first order Markov Chain is a finite or countably infinite
sequence of events {E1, E2, … } over discrete time points, where Pij = P(Ej | Ei), and at any time the future behavior of the process is based solely on the current state
A Markov Model (MM) is a graph with m vertices or states, S, and directed arcs, A, such that:
S ={N1,N2, …, Nm}, and A = {Lij | i 1, 2, …, m, j 1, 2, …, m} and Each arc,
Lij = <Ni,Nj> is labeled with a transition probability Pij = P(Nj | Ni).
10/24/08 17
Problem with Markov Chains
The required structure of the MC may not be certain at the model construction time.
As the real world being modeled by the MC changes, so should the structure of the MC.
Not scalable – grows linearly as number of events. Our solution:
Extensible Markov Model (EMM) Cluster real world events Allow Markov chain to grow and shrink dynamically
10/24/08 18
Outline
Introduction EMM Overview EMM Applications to Anomaly
Detection Future Work
10/24/08 19
Extensible Markov Model (EMM)
Time Varying Discrete First Order Markov Model Nodes are clusters of real world states. Learning continues during application phase. Learning:
Transition probabilities between nodes Node labels (centroid/medoid of cluster) Nodes are added and removed as data arrives
10/24/08 20
Related Work Splitting Nodes in HMMs
Create new states by splitting an existing state M.J. Black and Y. Yacoob,”Recognizing facial expressions in image sequences using local
parameterized models of image motion”, Int. Journal of Computer Vision, 25(1), 1997, 23-48. Dynamic Markov Modeling
States and transitions are cloned G. V. Cormack, R. N. S. Horspool. “Data compression using dynamic Markov Modeling,” The
Creates new states if the input data has never been seen in the model, and transition probabilities are adjusted
Dani Goldberg, Maja J Mataric. “Coordinating mobile robot group behavior using a model of interaction dynamics,” Proceedings, the Third International Conference on Autonomous Agents (agents ’99), Seattle, Washington
10/24/08 21
EMM vs AMMOur proposed EMM model is similar to AMM, but is more flexible: EMM continues to learn during the application phase. The EMM is a generic incremental model whose nodes can have any
kind of representatives. State matching is determined using a clustering technique. EMM not only allows the creation of new nodes, but deletion (or
merging) of existing nodes. This allows the EMM model to “forget” old information which may not be relevant in the future. It also allows the EMM to adapt to any main memory constraints for large scale datasets.
EMM performs one scan of data and therefore is suitable for online data processing.
10/24/08 22
EMM
Extensible Markov Model (EMM): at any time t, EMM consists of an MM and algorithms to modify it, where algorithms include:
EMMSim, which defines a technique for matching between input data at time t + 1 and existing states in the MM at time t.
EMMIncrement algorithm, which updates MM at time t + 1 given the MM at time t and classification measure result at time t + 1.
Additional algorithms may be added to modify the model or for applications.
10/24/08 23
EMMSim
Find closest node to incoming event. If none “close” create new node Labeling of cluster is centroid/medoid of
members in cluster Problem
Nearest Neighbhor O(n) BIRCH O(lg n)
• Requires second phase to recluster initial
10/24/08 24
EMMIncrement
<18,10,3,3,1,0,0><18,10,3,3,1,0,0>
<17,10,2,3,1,0,0><17,10,2,3,1,0,0>
<16,9,2,3,1,0,0><16,9,2,3,1,0,0>
<14,8,2,3,1,0,0><14,8,2,3,1,0,0>
<14,8,2,3,0,0,0><14,8,2,3,0,0,0>
<18,10,3,3,1,1,0.><18,10,3,3,1,1,0.>
1/3
N1
N2
2/3
N3
1/11/3
N1
N2
2/3
1/1
N3
1/1
1/2
1/3
N1
N2
2/3 1/2
1/2
N3
1/1
2/3
1/3
N1
N2
N1
2/21/1
N1
1
10/24/08 25
EMMDecrement
N2
N1 N3
N5 N6
2/2
1/3
1/3
1/3
1/2
N1 N3
N5 N6
1/61/6
1/6
1/31/3
1/3Delete N2
10/24/08 26
EMM Advantages
Dynamic Adaptable Use of clustering Learns rare event Scalable:
Growth of EMM is not linear on size of data. Hierarchical feature of EMM
Creation/evaluation quasi-real time Distributed / Hierarchical extensions
10/24/08 27
Growth of EMM
0
100
200
300
400
500
600
700
800
1 80 159
238
317
396
475
554
633
712
791
870
949
1028
1107
1186
1265
1344
1423
1502
number of input data (total 1574)
num
ber o
f sta
te in
mod
el threshold 0.994
threshold 0.995
threshold 0.996
threshold 0.997
threshold 0.998
Servent Data
10/24/08 28
EMM Performance – Growth Rate
Data SimThreshold
0.99 0.992 0.994 0.996 0.998
Serwent
Jaccrd 156 190 268 389 667Dice 72 92 123 191 389
Cosine 11 14 19 31 61Ovrlap 2 2 3 3 4
Ouse
Jaccrd 56 66 81 105 162Dice 40 43 52 66 105
Cosine 6 8 10 13 24Ovrlap 1 1 1 1 1
10/24/08 29
EMM Performance – Growth Rate
Minnesota Traffic Data
10/24/08 30
Outline
Introduction EMM Overview EMM Applications to Anomaly
Detection Future Work
10/24/08 31
Datasets/Anomalies MnDot – Minnesota Department of Transportation
Automobile Accident
Ouse and Serwent – River flow data from England Flood Drought
By learning what is normal, the model can predict what is not
Normal is based on likelihood of occurrence Use EMM to build model of behavior We view a rare event as:
Unusual event Transition between events states which does
not frequently occur. Base rare event detection on determining events
or transitions between events that do not frequently occur.
Continue learning
10/24/08 34
EMMRare
EMMRare algorithm indicates if the current input event is rare. Using a threshold occurrence percentage, the input event is determined to be rare if either of the following occurs: The frequency of the node at time t+1 is below
this threshold The updated transition probability of the MC
transition from node at time t to the node at t+1 is below the threshold
10/24/08 36
Determining Rare
Occurrence Frequency (OFc) of a node Nc :
OFc =
Normalized Transition Probability (NTPmn), from one state, Nm, to another, Nn :
NTPmn =
c ii
CN CN
mn ii
CL CN
10/24/08 37
EMMRareGiven:
• Rule#1: CNi <= thCN
• Rule#2: CLij <= thCL
• Rule#3: OFc <= thOF
• Rule#4: NTPmn <= thNTP
Input: Gt: EMM at time t
i: Current state at time t R= {R1, R2,…,RN}: A set of rules
Output: At: Boolean alarm at time t
Algorithm: At =
1 Ri = True
0 Ri = False
10/24/08 38
Rare Event in Cisco Data
10/24/08 39
Problem: Mitigate false alarm rate while maintaining a high detection rate.
Methodology: Historic feedbacks can be used as a free resource to take out some
possibly safe anomalies Combine anomaly detection model and user’s feedbacks. Risk level index