Abstract
Diabetes is among the socially significant diseases, which leads to high costs for the diagnosis and treatment of diabetes. Diagnosis and treatment of diabetes is currently one of the important tasks in medicine at the present stage of development of medical services. An important direction in the development of medical services for the population is the development and implementation of various problem-oriented information systems. Similar systems developed earlier did not cover the entire amount of heterogeneous information that is collected when diagnosing and prescribing the course of diabetes treatment, nor did they use technologies and cloud services as tools for Big Data. In this article, let’s make use of the predictive analytic to forecast and categorize the type of diabetes which offers an effective method for treating and curing patients at a reduced cost, with improved results such as affordability and availability. An information system platform has been developed and configured to manage the Hadoop cluster, as well as a non-relational database that uses and processes unstructured data in various formats. All experimental research, development of methods and algorithms, as well as solving computational problems were implemented using software languages for application development. The novelty lies in the research of distributed computing models that provide efficient execution of developed algorithms using the conceptual model of the processes of search, extraction and analysis of unstructured data in large data sets. The practical implementation of algorithms was carried out on the basis of methods of object-oriented programming and object-oriented databases