||A Proposal of a Privacy-preserving Questionnaire by Non-deterministic Information and Its Analysis
Tsumoto, Shusaku ,
Nakata, Michinori ,
Sakai, HiroshiLiu, Chenxi
2016 IEEE International Conference on Big Data (Big Data)
1965 , 2016-12-06 , IEEE
We focus on a questionnaire consisting of three-choice question or multiple-choice question, and propose a privacy-preserving questionnaire by non-deterministic information. Each respondent usually answers one choice from the multiple choices, and each choice is stored as a tuple in a table data. The organizer of this questionnaire analyzes the table data set, and obtains rules and the tendency. If this table data set contains personal information, the organizer needs to employ the analytical procedures with the privacy-preserving functionality. In this paper, we propose a new framework that each respondent intentionally answers non-deterministic information instead of deterministic information. For example, he answers ‘either A, B, or C’ instead of the actual choice A, and he intentionally dilutes his choice. This may be the similar concept on the k-anonymity. Non-deterministic information will be desirable for preserving each respondent's information. We follow the framework of Rough Non-deterministic Information Analysis (RNIA), and apply RNIA to the privacy-preserving questionnaire by non-deterministic information. In the current data mining algorithms, the tuples with non-deterministic information may be removed based on the data cleaning process. However, RNIA can handle such tuples as well as the tuples with deterministic information. By using RNIA, we can consider new types of privacy-preserving questionnaire.
2016 IEEE International Conference on Big Data, December 5-8, 2016, Washington DC, USA