Ethics and Data Mining in Biomedical Engineering

Abstract

Data mining is a generic term given to the process of analysing data, usually high volumes of data contained in large databases, in order to discover previously unknown patterns and trends. Data mining utilizes and combines methods from statistics, machine learning, pattern recognition and database management. Typical data mining tasks involve detecting data subsets that are similar in some way, unusual or anomalous or have features that are associated or dependent. Although data mining is not traditionally focused on the development of predictive models that generalise known patterns to new (unseen) data, this is often an extremely valuable way of verifying the efficacy of the derived models. That is, if these models accurately predict unseen data, then is more likely that these models truly represent the underlying patterns in the data rather than being purely by-chance occurrences. In addition, in biomedical applications the purpose of the data mining is often to better understand the patterns of disease so that improved diagnoses, prognoses and treatments can be developed in the future.