Hospital Data Interpretation: A Self-Organizing Map Approach

Abstract

It is not feasible to attempt to interpret complex health care data with large dimensions using typical 2D or 3D charts, diagrams, or graphs. It is helpful to be able to correlate the dimensions against one another to discover new patterns and obtain fresh knowledge. Effective interpretation of the statistical data, collected from health care centers, helps physicians and clinicians to improve their efficiency and the quality of care. This work has used different types of Self-Organizing Maps (SOM) in order to provide visual interpretability of the collected data to the hospital administration. Using the approach presented in this work, existing correlations among different attributes of collected data can be discovered and utilized to uncover hidden patterns. We illustrate how Self-Organizing Maps can be effectively used in interpretation of health care and similar data.