Approaches for knowledge discovery in databases

The paper presents and discusses issues related to the analysis and interpretation of medical data thus allowing knowledge discovery in medical databases. Knowledge in a knowledge-based system for decision support is generally acquired from experts and from the literature. We have shown that knowledge can also be effectively extracted from a database of patient observations and from interpretation of those observations. The resulting system is more accurate in our case, especially if it is intended to operate as a decision support system in the same clinical setting as the one providing the original data. The paper shows how retrospectively collected data could be utilized for the purpose of knowledge extraction. Traditional data analysis approaches were used to analyze a database on liver diseases