A hybridized rough set and bat-inspired algorithm for knowledge inferencing in the diagnosis of chronic liver disease

Healthcare informatics data is proliferating, and analyzing this data is a challenging issue as it requires multiple levels of extraction of knowledge for decision making. Knowledge discovery of databases is a solution to this end. Nevertheless, healthcare data contains uncertainties, and so there is a need for computational intelligence techniques to process such uncertainties while considering feature selection, classification, clustering, and decision rule generation. The rough set is a relatively new technique towards decision rule generation without considering any additional information. On the other hand, bio-inspired computing techniques are widely used for optimization and feature selection. Primarily, bio-inspired computing uses a minimum number of features to classify a system. Therefore, the integration of rough set and bio-inspired computing leads to optimal rule generation. Keeping it in mind, in this paper, we integrate a rough set and bat algorithm to foster knowledge. At the initial phase, the bat algorithm is employed to identify the chief features that affect the decisions. Further, decision rules are generated using these selected features. It, in turn, helps to diagnose a disease at an early stage. The objective is not to replace a physician but to give an alternative opinion to the physician. It is believed that the proposed system can be used as a tool for the prevention and detection of malignancy of various communicable and non-communicable diseases. Simultaneously, it paves the way for efficient healthcare informatics. A case study on chronic liver disease is considered for analyzing the proposed model. Further, the obtained results are compared with hybridized decision tree algorithms and found significantly better.

Automated summary

Healthcare informatics data is challenging to analyze due to its complexity. This paper proposes an integration of rough set and bat algorithm to generate optimal decision rules and aid in the early diagnosis of diseases, providing an alternative opinion to physicians. The model shows promising results in preventing and detecting various diseases.