Developing Machine Learning Techniques for Diseases Prediction in Big Data

Abstract

A big-data revolution is underway in health care. Start with a much-improved amount of knowledge. Over the last decade, pharmaceutical organizations have been aggregating years of analysis and community data into medical databases, while payors and providers have digitized their patient records. The significant growth of big data in biomedical, healthcare teams, and scientific analysis of medical data advantages early disease detection, patient care, and community services. However, the analysis efficiency decreased when the quality of medical data is inadequate. Further, various sections exhibit unique features of certain regional diseases, which may reduce the prediction of disease outbreaks. In this study, we recommend conventional machine learning algorithms for investigation using feature uprooting expertise of Deep Neural Network studies, and we are stimulated by transfer learning methodology to this. The main aim of this approach is to examine the need for a deep Learning based security algorithm for massive data analysis. To overcome the problem of incomplete data, we use Deep Belief Networks with Disease Risk Prediction Algorithm (DBN-DRP) for structured and Unstructured data from Hospital. The DBN-DRP algorithm provides outstanding performance in big data classification than the previous methods. Compare to Existed Classification algorithms our proposed algorithms will offer better performance in classification.

Keywords: Deep learning, Deep Belief Networks, Disease Risk Prediction Algorithm, Deep Neural Network.