Lebesgue Prevosti Class Balanced Neural Turing Machine Data Classification For Healcathre Data Analytics

Abstract

Medical data refers to the health-related information associated with routine patient care and clinical trial program. Data mining has a phenomenal perspective for healthcare services owing to the mushrooming increase in electronic health records. Digitalisation and innovation of new mechanisms minimise human efforts and make data straightforwardly significant. Machine Learning (ML) technique is used in the healthcare domain to diagnose different diseases. Also, with the aid of classification techniques and healthcare data, significant disease diagnoses are said to be ensured. Therefore, feature selection as a key element is very significant in ML-assisted diagnosis. So far has designed several feature selection methods. However, inspecting the accuracy has developed hardly any observations and time complexity with rational feature selection. To address class-imbalanced data without fine-tuning, utilising traditional feature selection models for classification can only be done with a smooth process. Lebesgue Prevosti Class-balanced Neural Turning Machine Data Classification (LPC-NTMDC) proposed for disease diagnosis to address this issue. The LPC-NTMDC method has split into two parts, namely, feature selection and classification. First, the Lebesgue Prevosti Rationality-based Feature selection algorithm proposed to carry out rational sparse feature selection applied to complex class-imbalanced data. Following this, to evaluate the rationality of feature selection has employed a recurrence-based scale represents a pipeline to select inherent features considering both rationality and class imbalance. Second, with the inherent selected features, Neural Turing Machine Data Classification-based disease diagnosis is designed. Here with Fuzzy feature matching capabilities, significant medical data classification is made for disease diagnosis. Experimental results on the cardiovascular disease dataset have demonstrated the feasibility of our method. Furthermore, the experimental results stipulate that the proposed LPC-NTMDC method outperforms the state-of-the-art methods regarding classification time, accuracy, false positive rate and space complexity.