Centralized Large Data Set Storage System with File Classification Technique

To furnish effective rules determining the format and transmission for securing and privatizing k-Nearest Neighbor (k-NN) search, when the data is dealt between individual parties who want to cooperatively figure out the answers without disclosing either of their individual data. To develop the k nearest neighbors (KNN) from one dataset, of each and every detail in other dataset, the special and multimedia databases are useful tools in the kNN join. K-NN can be applied for both classification and regression predictive problems. Nevertheless, it is more widely used in classification problems as per the industry. To verify any technique we usually look at three important aspects repose to read output, Calculation time, Prognostic Power. We give a k-NN concealment model for saving the privacy of the patients in a cloud based e-healthcare system as the sensitive information is needed to be sustained in a confidential manner and not to be exposed to public users other than the medicos. However, collection of data’s are used for classification might have sensitive data, directly outsourcing them to any Distributed servers inevitably raise privacy concerns. The proposal of a practical concealment K-Nearest Neighbor classification scheme over bulk datasets using map reduce technique that can be efficaciously outsourced to HDFS servers is enforced.