Intelligent Information Structure Investigation in Biomedical Databases: The Breast Cancer Diagnosis Problem

V. Bevilacqua, G. Mastronardi, and F. Menolascina (Italy)

Keywords

Artificial Neural Networks, Breast Cancer Diagnosis,Statistical Analysis, DataMining, Simulated Annealing,KMean clustering.

Abstract

This paper presents a system for investigating data structure in a biomedical database. AI tools were used to explore complex space of data that similar datasets show. The technique described aimed to analyse data contained in Wisconsin Breast Cancer Database (WBCD)[1][2][3], results were therefore employed in Intelligent System designing. High dimensions and information redundancy characterise this database. The employment of statistical analysis techniques returned relevant basic information; even though their employment in dataset with more complex information distributions has highlighted important limits. In this paper we describe an intelligent pattern discovery method that uses Artificial Neural Networks (ANN) and other intelligent techniques to explore the space of parameters and to acquire useful knowledge for breast cancer diagnosis. Basic results obtained by ANNs were supported by Data-Mining techniques and Intelligent System validation. Potentialities of similar methods in biomedical databases analysis have been proofed in real world applications with competitive results.

Important Links:



Go Back