Proper

Abstract

Due to the technological development in health care informatics, digitalizing health records and telemedicine has resulted in rapid growth in the production of enormous amount of health data which clutches multifaceted information relating to patients and their medical circumstances. To extract and discover the useful information from outsized dataset data mining techniques uses which is applied on medicinal data to enhance the services of the health care division. Outlier detection is one of the data mining techniques which can be used in various field of medical division to analyze outsized dataset but the conventional outlier detection is not much efficient for large scale and high dimensional health care data. This paper proposes a hybrid outlier detection method namely ID3 with KNN and GA to develop effective outlier detection. The proposed ID3-KNN&GA agrees to the case categorization quality character (CCQC) with the medical quality evaluation model and utilizes the attribute overlapping rate (AOR) algorithm for data classification and dimensionality reduction. To calculate the performance of the pruning operations in ID3-KNN & GA, we perform widespread experiments on accuracy and specificity parameters. The experiment consequences show that the ID3-KNN&GA method outperforms the k-nearest neighbor (KNN) in terms of the accuracy and efficiency.