Comparison of K-Means clustering and statistical outliers in reducing medical datasets

Data reduction is a process of reducing the datasets in volume, almost used in all real time applications. Although there are several techniques available, many researchers have used K-Means clustering in reducing the datasets. In this paper, three different methods were used to replace missing values with mean, median and a predicted score; the cleaned datasets were reduced using K-Means clustering and Statistical Outlier detection. This research work compares the data reduction percentage performed by K-Means and Statistical Outliers for all the three methods of imputation. The experimental result proves that, the reduction rate of outliers is less than K-Means clustering.