Applications in Engineering
Data mining has been successfully applied in domains such as biology (DNA microarrays), spatial analysis, web mining, business and many others. Data mining as already been used in engineering as well. Examples of application include oil production prediction, traffic pattern recognition, composite joint behavior and joint damage assessment. However, all of these contributions use data mining as a predictive tool. In previous work at EPFL, data mining techniques have been applied to models to extract knowledge. Although the information obtained has been useful, it is limited to confirming relationships between parameters and their importance. For example, correlation and principal component analysis help indicate independent parameters. Decision tree analysis is capable of finding parameters that are able to separate models. However, no information related to the space of possible models is provided. Furthermore, no guidelines are provided to use the results and the limitations of such a methodology have not been clearly evaluated.
For more details about possible applications of data mining in the field of engineering and more particularly civil engineering, you can have a look at my research interests.
