Secure Rule Generation and Deep Computational Feature Learning model for Cloud Based Big Data

In the work of big data analysis and its mining, feature learning plays a very important and fundamental role. The big data has various properties including volume, variety and velocity. Such properties become the challenging problems in the process of feature learning. Aiming to these problems, deep computational learning model is proposed here. This model is extended with the algorithm named as privacy preserving high order back propagation algorithm. The security of information is maintained during the deep computational over cloud big data, with the BGV fully homomorphism encryption scheme. In addition, rule generation method for frequent pattern generation is used with FP-Growth algorithm. The experimental results prove that the proposed system performs effective and efficient feature learning over cloud based big data.