The explosive growth in the volume of data and the growing number of disparate data sources is bringing enormous opportunities and challenges to many research communities. In the biomedical domain, the challenge of knowledge discovery from diverse and heterogeneous biomedical data sources in order to make knowledge and concepts sharable over applications/experiments and reusable for several purposes is both complex and crucial. Opportunities arise by the simple act of connecting different facts and points of view that have been created for one purpose, but that in light of subsequent information can be reused in a quite different context, to form new concepts or hypotheses. However such interactions cannot be determined in advance - for one thing, there may be more or fewer problem dimensions involved in a process than were known when the process initially started. Modelling of such processes is a challenging task but is one with practical applications in many disciplines. Identifying these data interactions, learning about them, extracting knowledge, and building a reusable knowledge base that applies leading artificial intelligence and soft-computing methods will guide future research and practice and is at the core of this research. The novel Ontology Driven Knowledge Discovery framework (ODKD) developed in this research provides a means of describing and representing evolving knowledge, managing shared knowledge, integrating data mining tools and algorithms, and enabling semantically rich knowledge discovery. The ODKD suite of tools implements a framework able to integrate the evolving ontology meta-knowledge model and methodology to provide a more holistic view of the knowledge discovery in databases process than previously possible. In this thesis the functional capabilities of the tools and the appropriateness of the conceptual structures are demonstrated and evaluated in the context of a biomedical application case study.