Diabetes mellitus is a serious health problem affecting the entire population all over the world for many decades. It is a group of metabolic disorder characterized by chronic disease which occurs due to high blood sugar, unhealthy foods, lack of physical activity and also hereditary. The sorts of diabetes mellitus are type1, type2 and gestational diabetes. The type1 appears during childhood and type2 diabetes develop at any age, mostly affects older than 40. The gestational diabetes occurs for pregnant women. According to the statistical report of WHO 79% of deaths occurred in people under the age of 60, due to diabetes. With a specific end goal to deal with the vast volume, speed, assortment, veracity and estimation of information a scalable environment is needed. Cloud computing is an interesting computing model suitable for accommodating huge volume of dynamic data. To overcome the data handling problems this work focused on Hadoop framework along with clustering technique. This work also predicts the occurrence of diabetes under various circumstances which is more useful for the human. This paper also compares the efficiency of two different clustering techniques suitable for the environment. The predicted result is used to diagnose which age group and gender are mostly affected by diabetes. Further some of the attributes such as hyper tension and work nature are also taken into consideration for analysis.
Keywords: Cloud computing; Clustering techniques; Diabetes mellitus; Dynamic data; Hadoop.