Heart Diseases Prediction Using Unsupervised K-means Clustering Data Mining Technique – An Experimental Approach

Data mining is an advanced and interactive approach to explore and obtain the useful patterns of information from the large, undiscovered and distant data warehouses. There is enormous raw data available in the form of text, images, and videos in health care management systems. Data is stored in the medical databases after every microsecond. So there is an open confront to extract these hidden and undiscovered patterns to utilize for diagnosing systems in medical industry. Data mining techniques in medical science plays an imperative role to devise predictive systems for efficient diagnosis. In this work a heart diseases predictive system is devised using k-means clustering mining technique to form the high quality and efficient clusters. The primary goal of k-means algorithm is to form groups based on the identical data objects. The model which shows better performance results based on the size of clusters selected, is adopted and proposed in this work. Numbers of clusters i.e. k is varied to select the best clustering size in the proposed system. To perform experimental efforts, weka simulation tool is employed.