An Optimum Model for the Retrieval of Missing Values for Data Cleansing using Regression Analysis

An important aspect of the data mining is the pre-processing of the data. Pre-processing of the data is important because real world data is susceptible to inconsistencies, noise and missing values. Such a data cannot be used in data mining as that would produce highly inadequate results .There are basically two methods through which we can remove the problem of the missing values the first one is to ignore the data set with the missing value the second one is to predict those values. Prediction can be made based on assuming the continuity of the data or giving them some suitable value based on previous knowledge .In this paper our focus is on providing an adequate method to fill those missing values by predicting a suitable value by comparing and choosing a suitable regression method based on both the statistical and the subjective analysis of the graph from the various known regression method. General Terms Data Cleaning, Data warehousing, Regression techniques for data cleansing.