Modified Deviation Approach to Deal with Missing Attribute Values in Data Mining with different Percentage of Missing Values

Pallab KumarDey, Sripati Mukhopadhay
2013 International Journal of Computer Applications  
Information System having missing attribute values (in practical) hampers accurate estimation of Data Mining. If missing attribute values can be predicted in the pre-processing stage of data mining then it will help to improve the accuracy, and the existing data mining algorithms can also be applied based on complete data. In this work different type of methods available to handle incomplete information system have been discussed, and there after an algorithm has been proposed by which missing
more » ... ttribute values may be replaced with minimum complexity. It is shown that proposed algorithm is better by applying it on different sets of data with different percentage of missing values.
doi:10.5120/12734-9634 fatcat:4j4uholwyvcubhqvztjiqtdnzq