Question

Suppose that your local bank has a data mining system. The bank has been studying your debit card usage patterns. Noticing that you make many transactions at home renovation stores, the bank decides to contact you, offering information regarding their special loans for home improvements.

(a) Discuss how this may conflict with your right to privacy.

(b) Describe another situation in which you feel that data mining can infringe on your privacy.

(c) Describe a privacy-preserving data mining method that may allow the bank to perform customer pattern analysis without infringing on its customers’ right to privacy. (

d) What are some examples where data mining could be used to help society? Can you think of ways it could be used that may be detrimental to society?

Answer 1

(a) This is a major infringement on ones privacy as it is using our(the users) data to improve their business verticals and gaining profits out of it, the bank may also sell our information to other business and companies, this is a huge attack to our data privacy as other companies might build strategies and may also use our data for harmful activities, basically controlling us from all angles, data literacy and ethics should be maintained all over in any data mining task, user should have all permissions to protect his/her information

(b) A classic example where an analytical company used facebooks data to target voters in the U.S elections, this is highly unethical as with users personal information and demographic, voters can be manipulated and targeted to favour a certain candidate, which also an attack to the democracy priniciples, elections can be tampered and whoever has the correct demographic data can manipulate and rig elections, so this one example where data mining can go wrong.

(c)Since the primary task in data mining is to develop models about aggregated data, we can develop accurate models without accessing precise information in individual data records. For example, when building a decision-tree classifier, we can use training data in which the values of individual record shave been perturbed (i.e., data obscuration. The resulting data records look very different from the original records and the distribution of data values is also very different from the original distribution. The original distribution of a collection of distorted data values can be approximated using areconstruction algorithm. While it is not possible to accurately estimate original values in individual data records, Agrawal and Srikant [AS00] proposed a novel reconstruction procedure to accurately estimate the distribution of original data values. We can then build a classifier using the approximated values, thereby protecting the privacy of individuals. By using these reconstructed distributions, the resulting classifier will be the same as or very similar to that built with the original data.

(d)

One example where data mining can help society is through the provision of useful knowledge in theretail industry. It can help identify customer buying behaviors, discover customer shopping patternsand trends, improve the quality of customer service, achieve better customer retention andsatisfaction, design more effective goods transportation and distribution policies, and reduce the costof business. Another example where people can use data mining knowledge is to predict geneticdiseases. Using correlation analysis, one can identify co-occurring gene sequences and help predictvarious diseases.Alternatively, the results of data mining may be used in ways that could be considered detrimental tosociety. For example, suppose that the data mining of medical records found that most elderly patientswho receive hip replacement surgery do not survive for more than two years. Some medical insurancecompanies may refuse to pay for such surgery, knowing that it may not be successful for most cases.