Classification for the analysis of incomplete medical data

Abstract

Missing data is common in life. The processing of missing data is the key to classification. Therefore, using the existing reliable data set to complete the missing data is a common and necessary method. These methods have an important impact on dealing with the fuzziness and uncertainty in the data set. Therefore, it is necessary and effective to use accurate data and attribution methods to attribute missing data sets. This paper presents a new missing data classification method. Firstly, the center vector representing each class is calculated by using the training samples. The missing values are then estimated using the center of each class. By comparing the performance of three different interpolation methods in different test data sets, the final results show that the proposed method performs best in general.