Multi-view Analysis of Temporal Data

Multi-view data analysis approaches are attracting more and more attention recently due to the increasing availability of multi-source data in real-world applications, e.g. continuous monitoring of the operation of industrial assets (wind turbines, compressors, …) in the field generates a vast amount of data originating from a multitude of very diverse sources (operating parameters, environmental conditions, …). However, it is not always feasible and neither wise to approach data analytics for such applications by merging all the available data into a single data set.

In order to conduct meaningful analytics on complex data one should be able to consider and analyse separately different data representations (or views) provided by the different sources and take full advantage of the extracted knowledge by applying an appropriate integration analysis. Multi-view data analysis algorithms aim to solve these challenges by respecting the latent structure of the different data views. Dr. Tsiporkova will demonstrate the potential of multi-view analysis on real-world data by illustrating how it can be used for the integration of heterogeneous temporal data originating from multiple sources and for the establishment of a direct link between the operating modes and the prototypical performance of an industrial asset.