Rough set driven feature selection and rule based medical data classification approaches based on MapReduce computing framework

Abstract

The current study proposes an alternative strategy for managing huge and intricate datasets by integrating a number of information removal strategies, including Correlation-based Feature Selection (CFS), Best-First Search (BFS), and Dominance-based Rough Set Approach (DRSA). The goal of this learning is to improve the classifier's classification presentation by removing uncorrelated or unpredictable information values. The planned approach, dubbed CFS-DRSA, entails numerous stages. The operations are carried out sequentially, with two critical feature extraction techniques applied throughout the process's initial phases. Data reduction can be accomplished in the first phase by utilising both the CFS approach and the BFS algorithm. Second, a DRSA is used with a data selection technique to get the most optimal dataset possible for the circumstance. As a result, the investigation's primary objective is to identify a solution to the problem. Machine learning techniques can be used to increase classification precision while minimising calculation time. By including a variety of characteristics and volumes into the design, the experimental strategy was used to authenticate the planned technique. It was able to demonstrate the method's dependability and reliability using widely used assessment methodologies. When compared to other well-known methods, such as deep learning, this is very remarkable (DL). On the overall, the concept is advantageous because it has been demonstrated to aid the classifier in accurately re occurring a relevant result. When applied to incessant value datasets in which the information opinions do not contain any period info and are potentially inaccurate and unclear, the suggested model CFS-DRSA is effective. To validate the performance of the CFS-DRSA technique, a detailed experimental analysis is carried out and the experimental result highlights the betterment of the CFS-DRSA technique.