Abstract: This work aims at studying personal data analysis area, when confidentiality property of data is ensured. It is supposed that we are given partially critical social science data and prior to the submission of data to the public it is required to modify them so that confidential information is not disclosed, and that the analysis of these data did not differ from the analysis of raw data. Our work builds improved algorithms of class of classification and regression trees, which provide solution to the problem of generation of the so-called synthetic data. The new solution of generation takes into account the structure of the areas of privacy and is providing optimized tree replacement for synthetic data sets.

Keywords: classification, regression, data disclosure, synthetic data.

ACM Classification Keywords: H.1 Information Systems – Models and principles, I.2.0 Artificial intelligence.

Link:

Enhanced Cart Technologies in Partial Synthetic Data Generation

Levon Aslanyan, Vardan Topchyan

http://www.foibg.com/ijicp/vol01/ijicp01-03-p01.pdf