Imputation of medical records is a prime challenge when we deal with medical records. The imputed values affect classification of new medical records. This is because of this reason imputation gains its importance. In our present approach for imputation, we aim to carry imputation by first reducing dimensionality of medical records and then use the reduced dimension medical records to carry out imputation. The imputation is then followed by classification of medical records. We discuss a case study to show the process of imputation and classification of records using the proposed approach. The approach is demonstrated for categorical attributes in this paper and can be extended for numerical attributes also. The approach followed for imputation is class based clustering approach.
Unfortunately, ACM prohibits us from displaying non-influential references for this paper.
To see the full reference list, please visit http://dl.acm.org/citation.cfm?id=2905273.