Most of the women are affected by BC (Breast Cancer) which is one of the dreadful diseases in the entire world and considered as subsequent threatening reason of cancer death in women. The likelihood of death can be significantly reduced by means of early detection and prevention. Hybrid Bayesian frameworks were utilized previously for breast cancer prediction and handling missing values in patient’s characterization. WCP (Weight Canonical Polyadic) algorithms manage continuous missing values in the data by using least squares recursively. A main bottleneck in using WCP is the unfolding of multiple relationships of discovered modes (N). The complexity increases when the value of N is large. This paper uses imputations in attribute dependencies for enhancing BC detections. This work divides the dataset into discrete and continuous subsets where discrete fields are assigned values using BN (Bayesian Networks) followed by Tensor factorization on an integrated dataset using MFWCP (Mode Fuzzy Weight based Canonical Polyadic). The new dataset is created from full/missing value subsets for assigning values to fields with missing values. MFWCP operations result in operations where N value of WCP is greater than three. This third order is reduced to first order WCP by applying Khatri-Rao product. DT (Decision Trees), KNN (K-Nearest Neighbors) and ANFIS (Adaptive Neuro Fuzzy Inference System) classifiers are combined to classify BC and the proposed hybrid method is evaluated using defined performance measures enhanced imputation accuracy