Privacy Preserving Fuzzy Modeling for Secure Multiparty Computation

Many studies on privacy preserving of machine learning and data mining have been done in various methods by use of randomization techniques, cryptographic algorithms, anonymization methods, etc. Data encryption is one of typical approaches. However, its system requires both encryption and decryption for requests of client or user, so its complexity of computation is very high. Therefore, studies on secure computation using shared data are made to avoid secure risks being abused or leaked and to reduce computing cost. The secure multiparty computation (SMC) is one of these methods. So far, some studies have been done with SMC, but complex calculation processing such as machine learning has never proposed yet. In the previous paper, we proposed BP learning method for SMC on cloud computing system. In this paper, we propose learning method(Fuzzy modeling) of fuzzy inference system for SMC and prove the validity of it. Further, the performance of the proposed method is shown in numerical simulations.