Fuzzy rules-based Data Analytics and Machine Learning for Prognosis and Early Diagnosis of Coronary Heart Disease

Abstract

Globally, cardiovascular diseases stand as the primary cause of mortality. In response to the imperative to enhance operational efficiency and reduce expenses, healthcare organizations are currently undergoing a transformation. The incorporation of analytics into their IT strategy is vital for the successful execution of this transition. The approach involves consolidating data from various sources into a data lake, which is then leveraged with analytical models to revolutionize predictive analytics. The deployment of IoT-based predictive systems is aimed at diminishing mortality rates, particularly in the domain of coronary heart disease prognosis. However, the abundant and diverse nature of data across various disciplines poses significant challenges in terms of data analysis, extraction, management, and configuration within these large-scale data technologies and tools. In this context, a multi-level fuzzy rule generation approach is put forward to identify the characteristics necessary for heart disease prediction. These features are subsequently trained using an optimized recurrent neural network. Medical professionals assess and categorize the features into labeled classes based on the perceived risk. This categorization allows for early diagnosis and prompt treatment. In comparison to conventional systems, the proposed method demonstrates superior performance.