A COMPARISON OF TWO SINGLE IMPUTATION METHODS FOR HANDLING MISSING VALUES IN LARGE DATASET

Abstract

In real world, data may be incomplete, inconsistent or noisy. Missing values may occur due to several reasons. Data pre-processing is required in order to improve the efficiency of an algorithm. One of the challenging issues in data pre-processing is to handle the missing values in machine learning and data mining. There is a need for quality of data, thus it is ultimately important. To recover the solution of missing values the imputation techniques such as single, multiple and iterative imputations are there. The performance of the proposed algorithm has been compared with the other simple and efficient imputation methods. We compare Mean based Single Imputation (MI) and Standard Deviation Imputation (SDI) for effectiveness and improvement.
Key Words: Data mining, Pre-processing, Imputation, Mean Imputation.