INTERNATIONAL JOURNAL FOR INNOVATIVE RESEARCH IN MULTIDISCIPLINARY FIELD

Abstract: The Journey of data and its mining with analysis is endless, the basic types of data are numerical and character, other are hybrid. Usually, dataset or attributes with missing values confuses both the data mining and data analysis. It also affects the application of the result to novel data as well as concluding result. It also generated ambiguities in the datasets, need to treat before applying data mining, analysis or machine learning model. In order to recover missing values, earlier closest fit approaches are utilized during the data cleaning phase of pre-processing, resultant in excellent result with linear-numeric dataset. The present study is a comparison between three key closest fit approaches and their consequences.  It is significant review and comparative study of these approaches and their all-time applicability in the domain of data mining, analysis and machine learning.