Global, Local, and Transductive Model for Multiple Time-Series Analysis and Modelling

Time-series modelling and prediction have been very well researched by both the Statistical and Data Mining communities. However, the multiple time-series problem of modelling and predicting simultaneous movements of a collection of  time sensitive variables which are related to each other have received much less  attention eventhough the existence of dynamic relationships in multiple time-series data relating real world phenomena has been identified from previous studies.

The study focuses on the anaylsis and modelling of interactions in multiple time-series from a specific setting that take place continuously over time through different learning methodologies, to in the end perform simultaneous prediction. It is expected that various level of knowledge about the dynamics of the relationships in multiple time-series from a specific setting can be
acquired to finally construct a complete understanding about the underlying behaviour of the dynamic system under investigation.

IMMF Poster