SMART DISEASE PREDICTION USING CLUSTERING

Keywords

Abstract

Data mining is theprocess of extracting hidden prognosticinformation from large databases and is a powerful new technology with great potential. Mining useful knowledge from corpus of data has become a paramount application in many fields. There are many data mining algorithms for which performs operations like clustering, classification work on the data and provide crisp information for analysis. As these data are available through various channels into public domain, privacy for the owners of the data is significant. The healthcare industry contains large information, which is tedious to process by manual methods. In the existing system, they proposed a graph-based, semi-supervised learning algorithm called SHG-Health (Semi-supervised Heterogeneous Graph) for risk predictions. But there are two issues, performance issue and data source issue in the system. Medical datasets are often not balanced in their class labels. Most of the existing classification methods tend to perform poorly on dataset which is extremely imbalanced. Hence, an alternative method of modelling the objects is required .If we propose an efficient algorithm then the above said issues can be avoided. So in this paper we propose an algorithm for healthcare system to accurately predict the result from the large amount of data.