Data analytics for

Cyber security

-Understanding
sources of
Cybersecurity data-
Dr Fouzi Harrag
2024/2025

1
12/16/2024 Data Analytics for Cybersecurity
 End to End Opportunities for data collection

Sources of cybersecurity data

• Log Data
• Router Connectivity and Log Data
• Firewall Log Data
• Raw pay load data
• Network Topology data
Outline • User system data
• Other Datasets

Integrated use of multiple datasets

Summary of sources of Cybersecurity data

12/16/2024 Data Analytics for Cybersecurity 2

Sources of • Cyber threats often lead to loss of assets
Cybersecurity • Multitude of datasets can be harvested and used
to track these losses and origins of the attack
Data • This chapter is not about the data lost during
cyberattacks but the data that organizations can
scour from their networks to understand threats
better so that they can potentially prevent or
even predict future attacks

12/16/2024 Data Analytics for Cybersecurity 3

End to End Opportunities for data collection

A typical client server scenario

The information systems used to perform business
functions have a well-defined process spanning
over connected systems •A user connects to a system via an internet pipeline.
•The system has built-in application functionality important to run the
business function
•A return pipeline sends a response back to the user
•The functionality of the system allows the delivery of the information
commodity requested by the user

12/16/2024 Data Analytics for Cybersecurity 4

 Logical View
Logical View (a)

Request Request
Business Application
User Internet

Response Response

The logical view of the user requesting access to a business application can appear to be
fairly straightforward

Within this pipeline there could be several points through which the request and response
pass

Leading to several opportunities in the end-to-end process for data collection to help
understand when a cyber threat may occur in this process
12/16/2024 Data Analytics for Cybersecurity 5
 • User request on a network– follow a complex networking pipeline
• The user may have a firewall on their own system and the router through which
they send out the request
• This request can be filtered through the internet service provider
• lookups can be performed in the domain name system (DNS) and
• the data can be routed through multiple paths of routers, which are linked
Physical View through the routing table
• The request on the other side may again have to pass through the routers and firewalls
at multiple points in the system being accessed by the user
• There may be multiple intrusion detection systems (IDS) posted throughout the
systems to monitor the network flow for malicious activity
• This is just one example scenario; different network layouts will result in different
types of intermediate steps in this process of request and response, particularly based
on
• the type of response
• the type of network being used
• the type of organization of business applications
• the cloud infrastructure being used
• However, certain key components are always present that allow for multiple
opportunities to glean and scour for data related to potential cyber threats

Data Analytics for Cybersecurity 6

12/16/2024
Physical View
Request
User Domain
Firewall Name
Response Router Internet System
Service
Internet Provider Routing
Internet Table
Exchange
Enterprise Server
Routing
Internal IDS
Table
Systems
Internet
Service DB Files
Provider IDS
IDS
Router Switch
Firewall Request Firewall

IDS Web Server

Response Public Facing Systems
Business Application
12/16/2024 Data Analytics for Cybersecurity 7
Opportunities for Data Collection - to understand
potential threats.
Data collection begins at

• a user access point

• system functionality level, and
• commodity level (particularly if the data is being delivered)

Example – Questions to consider at the user level

• Who is the user?

• The psychology of the user, personality types, etc.
• What type of interface is being used by the user?
• Is there clear information about what is acceptable or not acceptable in the interface? (c)
What type of access system is being used? Is there access control for users?
• What data are available about the access pipeline, such as the type of network or cloud
being used?

12/16/2024 Data Analytics for Cybersecurity 8

Common types of cybersecurity data
Browser cache

Packet communication to derive Communication graphs

• Several common types of datasets
can be collected and evaluated,
including various types of log data
Key stroke logs
such as key stroke logs, web server
logs, and intrusion detection logs, to
name a few Packet sniffer log/tcpdmp

Router connectivity data/ router logs

Web server logs

Access control data

12/16/2024 Data Analytics for Cybersecurity 9

 • Cybersecurity-related data collection will vary across the type of networks, including
computer networks, sensor networks, or cyberphysical systems

• The method and level of data collection will also vary based on the application domains for
Sources of which the networks are being used and the important assets being protected

Cybersecurity • Social media businesses, such Facebook, are primarily user data driven, where the
revenue is based on providing access to user data and monitoring usage data

Data and • E-commerce businesses, such as Amazon, are usage and product delivery based

Variations • Portals, such as Yahoo, are again user data driven but more heavily reliant on
advertisements, which can target users based on what they see and use most often

• Cyberphysical systems, such as systems for monitoring and managing power grids, are
based on accurate functioning of physical systems and delivery of services to users over
these physical infrastructural elements

• The level of monitoring and management of data will vary with the level of prevention,
detection, or recovery expected in the domain
• Some domains have a high emphasis on prevention; others may have a high level of emphasis
on detection or recovery
• In all such cases, multiple types of datasets can be collected to provide intelligence on the
cyber threats, and user behaviors can be evaluated to prevent future threats or even identify
an insider propagating the threats

Data Analytics for Cybersecurity 10

12/16/2024
Log Data

Intrusion detection system (IDS) logs including alarms raised by IDS;

• Alerts are raised by matching any known signatures of malicious activities in the header
and payload data
• Logs analyze the packets based on malicious signatures and provide information on time
stamp, service used, protocol, source, and destination
• IDS can be placed at various points in a network, and multiple such datasets can be
collected and correlated

Key stroke logs

• To capture every key being pressed on a keyboard
• Capture actions such as copying materials to the clipboard or other interactions with the
user system

12/16/2024 Data Analytics for Cybersecurity 11

Log Data

Router connectivity data/ router logs;

• Routers not only provide route information but also all the raw IP addresses that pass through it
• IP addresses can be mapped to identify possible malware activity when data is sent to suspicious
geolocations in an unauthorized manner
• Router data can also be utilized to study and possibly identify traffic hijacking and bogus routes by
looking at historic route data stored in a knowledge base
Firewall logs
• Firewalls are typically designed to look at the header information in the data packets to match
against pre specified rule sets
• Firewalls differ from IDS since they are generally limited to header information screening whereas
IDS can look at the payload data as well and block connections with malicious signatures

12/16/2024 Data Analytics for Cybersecurity 12

 Raw Pay Load Data

Data over the There are several Payload data can be This data may be Payload data is Payload data can be
network contains privacy concerns in accessed only where encrypted, so its accessible through massive even for a
accessing this legally allowed and usefulness as raw packet sniffers such few minutes of data
The header payload data since users have provided data to be mined is as Wireshark, where capture
information, which this data is the permissions to limited the data dump of
stores data about actual content that access this data the traffic can be
source and is being sent which retrieved
destination among may be under strict
other things and access controls
the actual content
being transmitted,
referred to as
payload

12/16/2024 Data Analytics for Cybersecurity 13

 • To discover individual user’s behavior
• To detect presence of malwares in the payloads
• To detect other security threats based on the actual
content of the payload
• To identify threats based on signatures of malwares
that may be present in the payload
• For example, if a virus is embedded in a packet and
Raw Pay Load this virus has a known signature then this can be
Data - Uses captured by traditional intrusion detection system
rules
• Packets with malware embedded in them can be
detected using multiple mechanisms such as simple
keyword searches or complex regular expression
matches and flagged
• The traffic can be blocked or marked for further
analysis, such as using Snort alarms or Wireshark
coloring rules

Data Analytics for Cybersecurity 12/16/2024 14

Network Topology Data

A computer network can be represented as a Network traffic data dump can be used to Header data collected from a traffic dump file
graph in terms of the structure of the network and generate the communication graphs through Wireshark can be utilized to plot the
in terms of the communication taking place over communication between the source and
the networks destination IP addresses, which become the
vertices of each edge in the graph

12/16/2024 Data Analytics for Cybersecurity 15

Network
Topology
Data Preprocessing

Example
extraction of
communication
graph from
network traffic
Header data collected
from a traffic dump file
through Wireshark can
be utilized to plot the
communication
between the source
and destination IP
addresses, which
become the vertices of
each edge in the
graph.

12/16/2024 Data Analytics for Cybersecurity 16

Communications to Graphs

Communication data in the graph form,

• graph metrics can be computed such as node level metrics including centrality,
page rank and network level metrics such as diameter, density
• Based on the network properties future predictions can also be made about the
network evolution
The example illustrates one such task, in a sample traffic dump data.
• Data from network traffic can be collected through packet sniffers such as
Wireshark
• The data from the network traffic is preprocessed and this preprocessing will
change with the task being performed
12/16/2024 Data Analytics for Cybersecurity 17
 • Analysis by day of the week.
• The traffic data is sorted by the day of the week to get patterns
by day, such as all Mondays, all Tuesdays
• Compute the degree of each node by day of the week.
• Can be performed by specific date also
Example - • Can sort the IP addresses by their degrees across days of the week
Network and the top ones appear to be consistently present in the traffic
• Nodes with low degrees can also be identified
traffic to • In such a scenario it would be interesting to find a node which is
highly consistent as a high degree node to appear in the lower
Graph Analysis degree list indicating a shift in the traffic pattern
• Consider the bar chart of the degrees for each IP address across each
day of the week. We can observe that some IP addresses are
consistently higher across all days of the week, which is further
illustrated by the plot for IP1, IP2, and IP3 across all days of the week
• Degree of IP9 and IP7 seem to be higher on some days but lower on
other days
• Clarified by the plot for IP9, which shows Wednesday as a day where
IP9 has inconsistent behavior

Data Analytics for Cybersecurity 12/16/2024 18

 Data
Preprocessing

Consistently Central Nodes

High centrality
Monday
Monday Tuesday Wednesday Thursday Friday Average IP1 131102
IP1 131102 111662 125942 132846 130788 126468 IP2 67138
IP2 67138 71083 96073 68084 95038 79483.2 IP3 66826
IP3 66826 62722 83025 51658 88219 70490 IP4 55023
IP4 55023 47465 66643 64093 31606 52966
IP5 40120
IP5 40120 45089 35835 24626 48834 38900.8
IP6 34296
IP6 34296 29614 35447 40116 44573 36809.2
IP7 21365
IP7 21365 23757 34677 60977 14980 31151.2
IP8 42043
IP8 42043 42046 17620 34496 10305 29302
IP9 16360 16455 70181 13043 26243 28456.4 IP9 16360
IP10 43743 41197 26674 16867 9186 27533.4 IP10 43743
IP11 16642 13627 39616 46150 21056 27418.2 IP11 16642
IP12 30477 28288 30733 21490 18483 25894.2 IP12 30477
IP13 32864 33012 25680 21880 14237 25534.6 IP13 32864
IP14 22739 25507 19570 22308 31388 24302.4 IP14 22739
IP15 14875 15960 37694 13526 35597 23530.4 IP15 14875
IP16 20709 19761 20987 41284 14005 23349.2 IP16 20709
IP17 42364 23275 7512 36506 5389 23009.2 IP17 42364
IP18 24420 13220 42303 13150 20214 22661.4 IP18 24420
IP19 20462 20539 20749 9778 28250 19955.6 IP19 20462

19

12/16/2024
Data Analytics for Cybersecurity
 • Key features can be extracted to monitor unusual activities at the
individual system level
• Examples: active process resident memory usage, which is available
for all operating systems (OS) and allows for building a profile on the
normal memory usage of a process over time
• An abnormal spike in memory usage can be attributed to processing
a large volume of data
• Useful in detecting a potential insider threat, especially when
User System integrated with other user behavioral data from sensors monitoring
user stress levels or integrating with other log datasets

Data • CPU time utilization can be used for measuring system usage
• Several OS-specific features, such as kernel modules and changes in
registry values
• It is important to use multiple signatures over time from several of the
features to eliminate the regular spikes of day-to-day operations
• Key differentiator for a robust analysis where we do not simply rely on one
or two features but multiple features and their stable signatures (as
compared to historical data) to distinguish alerts
• Tools such as OSQuery and Snare can facilitate capture of these features

Data Analytics for Cybersecurity 12/16/2024 20

Key Features To Monitor Unusual Activities at the Individual System Level
Feature Name OS Specific
Active process name, Active process filesystem path, Active All OS
process ports and sockets, Active process file access, Active
process resident memory usage, Active process CPU time
utilization, Active process system calls, Active process priority
value, Active process owner and group information, Loaded
peripherals drivers, Key-store access Patterns

Loaded kernel modules Linux/Unix

Loaded Kernel Extensions Mac OS X
Change in registry values Windows
File System Journaling (meta-data) information All Major File System (NTFS, ext4,
HFS+)
Network Routing Tables All Major OS
Network Firewall rules All Major Firewall implementations

System level sensors (current, voltage in different bus inside PC, Almost all peripherals
CPU/GPU fan speed etc)

12/16/2024 Data Analytics for Cybersecurity 21

 • Access control data: These data can help better understand
usage of the assets that need to be protected. Role mining
from access control data can help shape and create better
and more robust roles
• Eye tracker data: A user’s behavior can be judged by the
interactions of the user with the system being used. One
such mode of input is the screen. Data collected from the
Other user’s eye gaze, captured through an eye tracker, can help
analyze the user’s level of engagement with a system and
Datasets user preferences or positioning important items on the
screen
• Vulnerability data: Software vulnerability is a defect in the
system (such as a software bug) that allows an attacker to
exploit the system and potentially pose a security threat.
Vulnerabilities can be investigated, and trends can be
discovered in various operating systems to determine levels
of strength or defense against cyberattacks

Data Analytics for Cybersecurity 12/16/2024 22

Example: NVD Datasets

National Vulnerability Database from the National Institute of Standards and Technology (NIST)

Trends can be analyzed for several years and across major releases for operating systems to
reinforce knowledge of choices for critical infrastructural or network projects

NVD is built on the concept of Common Vulnerabilities and Exposures (CVE), which is a dictionary of
publicly known vulnerabilities and exposures

CVEs allow the standardization of vulnerabilities across products around the world. NVD scores
every vulnerability using the Common Vulnerability Scoring System (CVSS)

CVSS is comprised of several submetrics, including (a) base, (b) temporal, and (c) environmental
metrics. Each of these metrics quantifies some type of feature of a vulnerability

12/16/2024 Data Analytics for Cybersecurity 23

 Cross-site Scripting vulnerability:
Data regarding the number of vulnerabilities pulled from NVD across 2006 to 2012
Comparing the occurrences of different types of vulnerabilities such as cross-site scripting and buffer overflow

WINDOWS MAC_OS_X LINUX ANDROID

14

BUFFER XSS
12
50

45
10
40

35
8
30

Number
6 25

20

4 15

10
2
5

0
0
2006 2007 2008 2009 2010 2011 2012
2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012
Year

12/16/2024 Data Analytics for Cybersecurity 24

Integrated Use of Multiple Datasets
• If multiple datasets result in similar types of anomalies, then the
credibility of labeling an anomaly is higher
• Example: to discover anomalies in network traffic data with a temporal,
spatial, and human behavioral perspective
• Studying how network traffic changes over time, which locations are the
sources, where is it headed, and how are people generating this traffic – all
these aspects become very critical in distinguishing the normal from the
abnormal in the domain of cybersecurity
• This requires shifting gears to view cybersecurity as a holistic people problem
rather than a hardened defense problem

12/16/2024 Data Analytics for Cybersecurity 25

 Computer networks evolve over time, and communication
patterns change over time

• Can we identify these key changes that are deviant from the normal changes in
a communication pattern and associate them with anomalies in the network
traffic?
Integrated Use of
As attacks may have a spatial pattern, sources and destinations in
Multiple Datasets: certain geolocations can be more important for monitoring and
Key Questions to preventing an attack

Consider • Can key geolocations that are sources of attacks, or key geolocations that are
destinations of attacks, be identified?
• Can IP spoofing be mitigated by looking at multiple data sources to supplement
the knowledge of a geospatial traffic pattern?

Any type of an attack has common underpinnings of how it is

carried out; this has not changed from physical security breaches
to computer security breaches
• Can this knowledge be leveraged to identify behavioral models of anomalies
where we can see patterns of misuse?

Data Analytics for Cybersecurity 12/16/2024 26

Summary of Sources of Cybersecurity Data
Source of cybersecurity
data Literature study examples Type of detection it can be used for
Simon 2007 , Gupta et al. 2016, Cai and Hao 2011, User behavior, malicious use to detect user
Keystroke logging Muzzamil et al. 2016 credentials

Deokar and Hazarnis 2012, Vaarandi and Podiņš 2010,

Quader and Janeja 2015, Quader and Janeja 2014, Chen et Association rule mining, human behavior modeling,
al. 2014, Janeja et al. 2014, Abad et al. 2003, Koike and log visualization, temporal analysis, anomaly
IDS log data Ohno 2004 detection

Router connectivity and Sklower 1991, Tsuchiya 1988, Geocoding Infosec 2013, Kim
log data Zetter Security 2013 , Jian 2007 Suspicious rerouting, traffic hijacking, bogus routes
Generate efficient rule sets, anomaly detection in
Firewall log data Golnabi et al. 2006, Abedin et al. 2010 policy rules

Wang and Stolfo 2004, Kim et al. 2014, Limmer and Malware detection, embedded malware, user
Raw payload data Dressler 2010, Parekh et al. 2006, Roy 2014 behavior
Massicotte et al. 2003, Nicosia 2013, Namayanja and Janeja Consistent and inconsistent nodes, time points
Network topology 2015 and 2017, corresponding to anomalous activity

User system data Stephens and Maloof 2014, Meigham 2016 User profiles, user behavior data, insider threats

Access control Data Vaidya et al. 2007, Mitra et al. 2016 Generate efficient access control roles
Browser security indicators, security cues, user
Eye tracker data Darwish and Bataineh 2012 behavior
Vulnerability data Frei et al. 2006 Vulnerability trend discovery
12/16/2024 Data Analytics for Cybersecurity 27
References
• Heron, Simon. "The rise and rise of the keyloggers." Network Security 2007.6 (2007): 4-6.

• Gupta, Haritabh, et al. "Deciphering Text from Touchscreen Key Taps." IFIP Annual Conference on Data and Applications Security and Privacy. Springer International Publishing,
2016.

• Cai, Liang, and Hao Chen. "TouchLogger: Inferring Keystrokes on Touch Screen from Smartphone Motion." HotSec 11 (2011): 9-9.

• Hussain, Muzammil, et al. "The rise of keyloggers on smartphones: A survey and insight into motion-based tap inference attacks." Pervasive and Mobile Computing 25 (2016): 1-25.

• Deokar, Bhagyashree, and Ambarish Hazarnis. "Intrusion Detection System using log files and reinforcement learning." International Journal of Computer Applications 45.19 (2012):
28-35.

• Vaarandi, Risto, and Kārlis Podiņš. "Network ids alert classification with frequent itemset mining and data clustering." 2010 International Conference on Network and Service
Management. IEEE, 2010.

• Quader, Faisal, Vandana Janeja, and Justin Stauffer. "Persistent threat pattern discovery." Intelligence and Security Informatics (ISI), 2015 IEEE International Conference on. IEEE,
2015.

• Chen Song, Janeja V., Human Perspective to Anomaly Detection for Cybersecurity, Journal of Intelligent Information Systems, Journal of Intelligent Information Systems, February
2014 (Accepted 2013) , Volume 42, Issue 1, pp 133-153

• Quader, Faisal; Janeja, Vandana, Computational Models to Capture Human Behavior in Cybersecurity Attacks Academy of Science and Engineering (ASE), USA, ©ASE 2014,
2014-06-16
•
• Janeja, Vandana P., et al. "B-dids: Mining anomalies in a Big-distributed Intrusion Detection System." Big Data (Big Data), 2014 IEEE International Conference on. IEEE, 2014.

• Abad, Cristina, et al. "Log correlation for intrusion detection: A proof of concept." Computer Security Applications Conference, 2003. Proceedings. 19th Annual. IEEE, 2003.

• Koike, Hideki, and Kazuhiro Ohno. "SnortView: visualization system of snort logs." Proceedings of the 2004 ACM workshop on Visualization and data mining for computer security.
ACM, 2004.

• Sklower, Keith. "A tree-based packet routing table for Berkeley unix." USENIX Winter. Vol. 1991. 1991.

• Tsuchiya, Paul F. "The Landmark Hierarchy: A new hierarchy for routing in very large networks." ACM SIGCOMM Computer Communication Review. Vol. 18. No. 4. ACM, 1988.

• Qiu, Jian, et al. "Detecting bogus BGP route information: Going beyond prefix hijacking." Security and Privacy in Communications Networks and the Workshops, 2007. SecureComm
2007. Third International Conference on. IEEE, 2007.

12/16/2024 Data Analytics for Cybersecurity 28

References
• Kim Zetter Security, WIRED, Someone’s Been Siphoning Data Through a Huge Security Hole in the Internet, [Link] 2013,
Last accessed 12/26/16
• Geocoding-Infosec: Robert Barnes, Infosec Institute, [Link] Geocoding Router Log Data, Aug, 8 2013
• Korosh Golnabi, Richard K. Min, Latifur Khan, Ehab Al-Shaer. Analysis of Firewall Policy Rules Using Data Mining Techniques[C]. Network Operations and Management
Symposium,2006. NOMS 2006. 10th IEEE/IFIP
• Abedin, Muhammad, et al. "Analysis of firewall policy rules using traffic mining techniques." International Journal of Internet Protocol Technology 5.1-2 (2010): 3-22.
• Wireshark, [Link]
• Wang K, Stolfo S. J. 2004. Anomalous Payload-based Network Intrusion Detection. In: Symposium on Recent Advances in Intrusion Detection, Sophia Antipolis, France.
•
• Sun-il Kim, William Edmonds, and Nnamdi Nwanze. 2014. On GPU accelerated tuning for a payload anomaly-based network intrusion detection scheme. In Proceedings of the 9th
Annual Cyber and Information Security Research Conference (CISR '14),
•
• Robert K. Abercrombie and J. Todd McDonald (Eds.). ACM, New York, NY, USA, 1-4. DOI=[Link]
•
• Tobias Limmer and Falko Dressler. 2010. Dialog-based payload aggregation for intrusion detection. In Proceedings of the 17th ACM conference on Computer and communications
security (CCS '10). ACM, New York, NY, USA, 708-710. DOI=[Link]
•
• Janak J. Parekh, Ke Wang, and Salvatore J. Stolfo. 2006. Privacy-preserving payload-based correlation for accurate malicious traffic detection. In Proceedings of the 2006
SIGCOMM workshop on Large-scale attack defense (LSAD '06). ACM, New York, NY, USA, 99-106. DOI=[Link]
•
• SNORT Rules Infographic [Link]
HMAC-SHA256&X-Amz-Credential=AKIAIXACIED2SPMSC7GA%2F20210316%2Fus-east-1%2Fs3%2Faws4_request&X-Amz-Date=20210316T191343Z&X-Amz-
Expires=172800&X-Amz-SignedHeaders=host&X-Amz-Signature=bcfc7d75d223ab40badd8bd9e89ded29cc98ed3896f0140f55320ee9bcdf1383, last accessed March 2020
• Cheok, Roy. "Wire shark: A Guide to Color My Packets Detecting Network Reconnaissance to Host Exploitation." GIAC certification paper (2014), SANS Institute Reading Room
• NodeXL: [Link]
• Nicosia, Vincenzo, et al. "Graph metrics for temporal networks." Temporal networks. Springer Berlin Heidelberg, 2013. 15-40.
• Namayanja, Josephine M., and Vandana P. Janeja. "Change detection in evolving computer networks: Changes in densification and diameter over time." 2015 IEEE International
Conference on Intelligence and Security Informatics (ISI).
• Namayanja, Josephine M., and Vandana P. Janeja. "Characterization of Evolving Networks for Cybersecurity." Information Fusion for Cyber-Security Analytics. Springer
International Publishing, 2017. 111-127.
• Massicotte, Frédéric, Tara Whalen, and Claude Bilodeau. "Network Mapping Tool for Real-Time Security Analysis." Real Time Intrusion Detection (2003).
• OSQuery. OSquery, last accessed, March 2020,. WWW: [Link] 2016.
• Snare. Snare, last accessed, March 2020, WWW: [Link]
• Stephens, Gregory D., and Marcus A. Maloof. "Insider threat detection." U.S. Patent No. 8,707,431. 22 Apr. 2014.
• Van Mieghem, Vincent. Masters Thesis Delft University, "Detecting malicious behaviour using system calls.", 2016
• Vaidya, Jaideep, Vijayalakshmi Atluri, and Qi Guo. "The role mining problem: finding a minimal descriptive set of roles." Proceedings of the 12th ACM symposium on Access control
models and technologies. ACM, 2007.
• Mitra, Barsha, et al. "A Survey of Role Mining." ACM Computing Surveys (CSUR) 48.4 (2016): 50.
• Darwish, A.; Bataineh, E., "Eye tracking analysis of browser security indicators," Computer Systems and Industrial Informatics (ICCSII), 2012 International Conference on , vol., no.,
pp.1,6, 18-20 Dec. 2012 doi: 10.1109/ICCSII.2012.6454330
• Frei, S., May, M., Fiedler, U. and Plattner, B., 2006, September. Large-scale vulnerability analysis. In Proceedings of the 2006 SIGCOMM workshop on Large-scale attack defense
(pp. 131-138). ACM.
• NIST. National institute of standards and technology: National vulnerability database. Accessed Sept, 2017, [Link]

12/16/2024 Data Analytics for Cybersecurity 29