Knowledge extraction using visualization of hemoglobin parameters to identify thalassemia

Abstract

The analysis of large amounts of data is better performed by humans when represented in a graphical format. Therefore, a new research area called the Visual Data Mining is being developed endeavoring to use the number crunching power of computers to prepare data for visualization, allied to the ability of humans to interpret data presented graphically.This work presents the results of applying a visual data mining tool, called FastMapDB to detect the behavioral pattern exhibited by a dataset of clinical information about hemoglobinopathies known as thalassemia. FastMapDB is a visual data mining tool that get tabular data stored in a relational database such as dates, numbers and texts, and by considering them as points in a multidimensional space, maps them to a three-dimensional space. The intuitive three-dimensional representation of objects enables a data analyst to see the behavior of the characteristics from abnormal forms of hemoglobin, highlighting the differences when compared to data from a group without alteration.