Visual approach for data mining on medical information databases using Fastmap algorithm

TL;DR

This paper introduces a visual data mining approach for large medical databases using FastMap for dimensionality reduction, emphasizing user-friendly graphical visualization for domain experts.

Contribution

It presents a novel software solution that visualizes multi-dimensional medical data with FastMap, focusing on ease of use for non-expert users in data mining.

Findings

Effective visualization of high-dimensional medical data

Facilitates domain experts' analysis without specialized data mining knowledge

Improves understanding of dependencies in medical datasets

Abstract

The rapid development of tools for acquisition and storage of information has lead to the formation of enormous medical databases. The large quantity of data definitely surpasses the abilities of humans for efficient usage without specialized tools for analysis. The situation is described as rich in data, but poor in information. In order to fill this growing gap, different approaches from the field of Data Mining are applied. These methods perform analysis of large sets of observed data in order to find new dependencies or concise representation of the data, which is more meaningful to humans. One of the possible approaches for discovery of dependencies is the visual approach, in which data is processed and visualized in a way suitable for analysis by a domain expert. This work proposes a visual approach, in which data is processed and visualized in a way suitable for analysis by a domain expert. We design and implement a software solution for visualization of multi-dimensional, classified medical data using the FastMap algorithm for graduate reduction of dimensions. The implementation of the graphical user interface is described in detail since it is the most important factor for the ease of use of these tools by non-professionals in data mining.