Big data is enormous, complex, comes at a faster rate and sometimes unstructured. Due to this aspect It is possible that an adversary can commit a digital crime without trace. Once you are able to secure the huge amount of Big data, how will you base your mode of segregation? Are you aware you might be filtering what should be actually what you are looking for?
This is an excellent question ! Threat actors use this to their advantage and hiding. If we sample the data in an inappropriate way to reduce the complexity we would definitely lose lot of evidence. For further reading, please check this excellent blog post on how not to sample
http://blogs.cisco.com/security/how-not-to-sample-network-traffic
@ Sanhank and Manesh, thank you for this useful links. I will go through them
Victor,
The challenge in this area of adversarial knowledge discovery is that you cannot assume in any analysis that the distributions are slowly varying. As soon as an adversary gains or assumes that data analysis is being carried out, there is a possibility that he will change the pattern to hide his behavior.
Look at the work of David Skillicorn for general overview on the subject or search under adversarial machine learning if you are into that specific area.
victor I think the idea is to start by mapping out potential criminal schemes, which is hard given the pace of technological evolution. Please check out the Accounting horizons journal....2015 latest edition for some write up on big data in audits...it has some good business perspective though not technical.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Statistical Analysis
- Data Analysis
- Dataset
- Big Data