Research to Protect Database by Shaking Random Sampling Interference (SRSI)
Tung-Shou Chen, Jeanne Chen, Yung-Ching Lin, Ying-Chih Tsai · 2009
Data mining is used widely by enterprises to mine hidden knowledge from databases. However, mined data containing sensitive trade secret could jeopardize the enterprise's competitive edge. In this paper, we proposed an anti-data mining concept to allow readable mined data that only contained unimportant information. The proposed shaking random sampling interference algorithm (SRSI) inserts interference data within a database to camouflage the real data. The scheme makes use of the data classification step in data mining to introduce interference data that has characteristics similar to the real data. Experimental results using four different classification algorithms showed that the interference data will decrease the accuracy of the database. The original database can be accurately recovered by using the correct parameters used in protecting the database.