MISSING VALUE IMPUTATION AND NORMALIZATION TECHNIQUES IN MYOCARDIAL INFARCTION

Abstract

Missing Data imputation is an important research topic in data mining. In general, real data contains missing values. The presence of the missing value in the data set has a major problem for precise prediction. The objective of this paper is to highlight possible improvement of existing algorithm for medical data. KNBP imputation method based on KNN and BPCA is proposed and evaluate MSE and RMSE estimates. Normalization is done by comparing three algorithms namely min-max normalization, Z-score and decimal scaling. The experiment is done with standard bench mark data and real time collected data. KNBP imputation method and Decimal Scaling Algorithm for Normalization got lower error rate.

Keywords

• Mean
• Hot Deck
• KNN
• BPCA
• KNBP
• Min-Max Algorithm
• Z-Score
• Decimal Scaling