An imputation measure for data imputation and disease classification of medical datasets
Shadi A Aljawarneh, Vangipuram Radhakrishna, Gunupudi Rajesh Kumar · AIP conference proceedings · 2019
Imputation of missing data values is an important pre-processing task for mining of medical data records. Application of data mining principles, techniques requires the dataset to be free from missing data values. In this paper, there are two contributions. One is the imputation measure for finding nearest optimal record for imputation and another is the algorithm for imputing missing data values. The proposed imputation function is extended by using our previous research in which a similarity measure for temporal pattern mining named as ASTRA is proposed. The measure ASTRA is modified suitably to serve as an imputation measure.