man analyzing data

Data Mining

Monitoring the operation of complex systems typically requires a large number of variables to be measured with high logging frequency. The resulting high data volumes must be properly structured in order to allow the analyst to draw relevant conclusions. Multivariate statistical methods like principal components and clustering are applied to identify which variables are really required for a sufficient usage monitoring and characterisation of typical usage profiles. In addition, the variance within a usage profile as well as among all relevant customer profiles is analyzed to provide a quantitative basis for accelerated testing and all further product development activities.