International Journal of Engineering and Technology

Diabetes mellitus is a chronic disease, also associated with an increased risk for heart disease and is emerging as a serious health challenge in India. It requires continuing medical care and patient self-management education to prevent severe problems and to reduce the risk of long-term problems. Healthcare industry stores a large amount of data which is not properly used to discover the hidden patterns and relationships. Disease diagnosis is one of the applications where data mining algorithms are proving successful results. In recent years, several researches have been conducted to develop intelligent clinical decision support system to help the physician in diagnosing the diabetes. This paper provides an introduction to the theory and highlights the importance of multivariate adaptive regression splines (MARS) in the disease diagnosis through the collected data for diabetes to develop an intelligent decision support system to help the physicians. MARS model obtained better accuracy with minimum number of predictors and outperformed by handling nonlinearities, missing data and interactions among predictors compared to neural network and other methods. The proposed approach is easily understandable, provides a better and faster model for diagnosing of diabetes patients.