Optimal Variables Identification and Statistical Mining Approach for Healthcare Data

B.V. Chowdary and Dr.Y. Radhika

The healthcare industries accumulate vast amount of health- care data that need to be mined to perceive hidden information for an adequate decision making process. In present day applications, many nursing homes handle medical information using health care information system. The major objective of this work is to design a creative Prediction System that provides analysis of disease using factual body parameters related data. Despite the maturing number of machine-learning (ML) algorithms that have been formed, still to implement them and provide the effectiveness and practicality is much desired. In this paper, feature variable’s importance identification is performed and a Heart Disease Prediction Model (HDPM) using statistical data mining approach is proposed. Using various medical profiles like - “age, blood pressure, asymptomatic chest pain, blood suger” etc., it will fore- cast the chance of patients obtaining any cardiovascular disease. It enables significant knowledge, e.g. various patterns present in data, relationships among medical factors, to be extracted. HDPM is adaptable, scalable, easily operated and expandable. The implementation is done using R framework. Experimental results proves its strength with im- proved prediction accuracy. Later, the comparison is performed for our proposed model (HDPM) with other significant previous works.

Volume 11 | 04-Special Issue

Pages: 1487-1495