A Novel Anti-Competitive Learning Neural Network Technique against Mining Knowledge from Databases

In this paper, we proposed an anti-competitive learning neural network scheme against mining of knowledge from databases. Neuron weights were trained by competitive learning in neural network and used with noise to harass the original database. The data mining process in anti-competitive learning will only allow data that contains unimportant knowledge to be mined. Experimental results showed that users can adjust neural weights to redirect harassment of the database to achieve the purpose of misleading illegal users and the mined data contained only unimportant knowledge.