Bias Protected Attributes Data Balancing using Map Reduce

In the discipline of Data Mining (DM), researchers occasionally disregard the necessity to maintain a balanced distribution on a dataset. This is problematic since maintaining a balanced distribution is essential. It is possible that the result of the process of categorizing things will be significantly influenced by this factor. Most classifiers are designed to function based on the fundamental premise that the distribution of the data is approximately balanced. Therefore, the efficacy of the classification technique has simply dropped, and it is required to address this issue to find a solution to the problem that has been presented. This research will construct a Fuzzy clustering using SMOT approach for the purpose of establishing labels for minority classes. The goal of this technique is to prevent the model from developing bias based on any of the protected attributes, such as gender, race, etc. This strategy makes advantage of parallel processing to provide a method for achieving a balance in the distribution of the dataset. Because of this, the classification findings will be enhanced in terms of the accuracy with which they categorize. The metrics of accuracy, sensitivity, and specificity are obtained from the various combinations of methods, and they are then compared to the various machine learning models that are currently in existence.