Knowledge extraction in large databases using adaptive strategies

Abstract

The general objective of this thesis is the development of an adaptive technique for extracting knowledge in large databases. Nowadays, technology allows storing huge volumes of information. For this reason, the availability of techniques that allow, as a first stage, analyzing that information and obtaining knowledge that can be expressed as classification rules, is of interest. However, the information available is expected to change and/or increase with time, and therefore, as a second stage, it would be relevant to adapt the knowledge acquired to the changes or variations affecting the original data set. The contribution of this thesis is focused on the definition of an adaptive technique that allows extracting knowledge from large databases using a dynamic model that can adapt to information changes, thus obtaining a data mining technique that can generate useful knowledge and produce results that the end user can exploit. The results of this research work can be applied to areas such as soil analysis, genetic analysis, biology, robotics, economy, medicine, plant failure detection, and mobile systems communications. In these cases, obtaining an optimal result is important, since this helps improve the quality of the decisions made after the process.