Probabilistic neural network based visual data mining for the healthcare sector

Abstract

BACKGROUND: The need for personalised care in the long-term management of patient health is paramount due to the variability in individual features and responses to specific medication. With the availability of large quantities of electronic patient records, big data analysis presents a valuable opportunity to gain insights into disease presentation and patient impact. OBJECTIVE: This study aims to utilise data science in the medical field to extract unknown information from databases, validate previously obtained data, and enhance personalised patient care. METHODS: An analytics suite is developed for monitoring patient health and treating cholesterol, thyroid, and diabetes disorders. This suite employs exploratory, predictive, and visual analytics to categorise patient data into multiple tiers and forecast related complication risk and treatment response. RESULTS: The study found that the analytics suite could successfully identify correlations between various biological indicators of patients and disorders. The suite also showcased potential in predicting health risks and responses to treatments. CONCLUSION: The analytics employed in this study suggest advanced methods of data analysis, which could serve as potential decision-making tools for healthcare providers. These methods might lead to improved treatment outcomes, contributing significantly to personalised patient care.