Paper Title
SVM Classification For Secured Encrypted Relational Data Over Cloud
Abstract
Nowadays web services and web users are rapidly increasing as a outcome a large information is produced.
Cloud technique is used to store such huge amount of data. With the utilization of users of cloud computing can get access
to store their data anywhere and any-time. As cloud servers are not trustworthy data needs to be an encrypted before
sending to cloud without losing the originality. For the use of real-time application like banking, medicine, scientific
research as well as between various governments agencies there is huge data need to be mined. Data mining applications
are commonly used for classification. Due to popularity of data classification different issues such as data cleaning, infinite
length, feature evolution and concept drift have arises. Various practical solutions have been provided under different
security models to solve these issues. In existing system classification problem had solved over encrypted data using
various privacy preserving classification protocol. Here the data stored was in an encrypted format using dataset record and
key size (k).This system gradually degrades system performance. To solve the above problem Support Vector Machine
(SVM) is utilized to classify the encrypted data under semi-honest model. The proposed system not only improved the
system performance but also proved to be a better methodology than existing methodologies under different parameter
settings.
Keywords— Data mining, Cloud computing, Privacy-preserving, Data encryption, Data decryption, SVM classifier.