AN IMPUTE MISSING VALUES USING IMPROVED WEIGHTED SMOTE FOR LIVER CELL IMBALANCED DATASET

In research field, the main problem is class imbalanced problem and impute missing attribute. In an imbalanced dataset occurs in various restraints when one of target classes has a small number of instances compare to other classes. Most of the oversampling m ethods may generate the wrong synthetic minority samples in some scenarios. To overcome this problem in the minority samples first identify the missing attribute data in correctly and learning the task easier. In this paper proposes the extension of Weight ed SMOTE called an Improved Weighted Synthetic Minority Oversampling Technique (IWSMOTE), which can overcome the problem of finding the missing attribute value of each samples for imbalanced liver cell dataset. The proposed algorithm evaluated based on exp erimental study. This algorithm compared against a existing SMOTE and Weighted SMOTE generalizations