Missing Data Imputation in Breast Cancer using Tensor and Baysian Networks

Missing value is an important problem in real-world dataset. Recentlysets of techniques have uses emerged for the imputation of missingdata. Handling this problem, especially in medical data whose controlis hard and costly, seems vital. Cancer refers to a disease in which agroup of cells shows uncontrolled growth (higher division than naturalcells), invasion (entering to adjacent tissues and disintegrating them)and sometimes metastasis. Without filling the missing data, the resultsof analysis might be doubtful. We tensor methods used to obtain themissing data. In this paper, we first get the missing values using theBayesian Networks and then use the tensor. Finally, the accuracy ofdisease diagnosis was estimated through support vector machine(SVM) classifier and RMSE measure. The results suggested that theproposed method is superior to other methods in term of value of error,accuracy, sensitivity and specificity.