BIG DATA ANALYZING FOR PREDICTING PATIENT FUTURE DISEASE USING NOVEL CGRBFN AND TFTCNN NEURAL METHODS
Electronic health records (EHR) have become more prevalent as a result of the quick development of information technology and Internet technology. This research paper focuses on introducing a novel deep learning approach for classifying the obtained information from feature selection process. The electronic health care records are considered as most important data that has to be protected from anonymous access. The novel CGRBFN model provides an additional cyber security system to product the anonymous access of EHR. The first stage of this research work talks more about removal technique involved for removing irrelevant information from collected EHR and also deals with noise removal techniques and null value entries. The Conjugate Gradient boosting architecture used in the proposed CGRBFN algorithm is chosen based on the attributes or features collected. The cryptographically approach based on block chain technology is used to improve the security of the proposed algorithm. The accuracy, precision, recall and f-measure of the EHR categorization analysis are made. The obtained accuracy for the two dataset MIMIC3 and CDSS are 0.939% and 0.9% respectively. The produced accuracy is more compared with other existing algorithms.