Towards an Artificial Intelligence System for Geographical Analysis of Health Data

Abstract

The complexity of modern scientific research requires advanced approaches to handle and analyse rich and dynamic data. Organizations such as hospitals, hold a great number of health datasets which may consist of many individual records. Artificial Intelligence methodologies incorporate approaches for knowledge retrieval and pattern discovery, which have been proven to be useful for data analysis in various disciplines. Decision trees methods belong to knowledge discovery methodologies and use computational algorithms for the extraction of patterns from data. This work describes the development of an autonomous Decision Support System (“Dth 1.0”) for the real-time analysis of health data with the use of decision trees. The proposed system uses a patient's dataset based on the patients’ symptoms and other relevant information and prepares reports about the importance of the characteristics that determine the number of patients of a specific disease. This work presents the basic concept of decision trees, describes the design of a tree-based system and uses a virtual database to illustrate the classification of patients in a hypothetical intra-hospital case study.