In the 'Dynamic Data Analytics through automatically Constructed Machine Learning Pipelines' project, the data dynamics of two focus projects are investigated.

One use case concerns degradation in Parkinson's patients with extremely long durations (weeks/months). The other use case involves energy management with extremely short timescales (minutes/hours).

For both cases, an optimized data analysis process is a necessary condition to predict: degradation (Parkinson's) or supply/demand (energy management).

Advanced big data techniques offer many possibilities for this. Both focus projects, in line with two top sectors, are logically connected by a general analysis platform for advanced big data techniques. This 'process platform' can eventually be applied to any theme where time series play an important role.

Projectpartners

Honda

Qualogy

TU Eindhoven

TU Delft