DocumentCode
2414482
Title
Sequential Data Fusion via Vector Spaces: Complex Modular Neural Network Approach
Author
Mandic, Danilo P. ; Goh, Su Lee ; Aihara, Kazuyuki
Author_Institution
Dept. of Electr. & Electron. Eng., Imperial Coll. London
fYear
2005
fDate
28-28 Sept. 2005
Firstpage
147
Lastpage
151
Abstract
A data fusion approach based on complex and hyper-complex vectors spaces is presented. The benefits of such an approach are highlighted and potential applications are identified. A case study on simultaneous forecasting of wind speed and direction in the complex domain, together with a distributed serial sensor fusion topology illustrate the potential of such an approach in real world applications
Keywords
forecasting theory; geophysics computing; neural nets; sensor fusion; time series; wind; distributed serial sensor fusion topology; hypercomplex vectors spaces; neural network; sequential data fusion; time series; vector spaces; wind speed direction; wind speed forecasting; Data engineering; Educational institutions; Frequency measurement; Industrial electronics; Neural networks; Particle measurements; Recurrent neural networks; Time measurement; Wind forecasting; Wind speed;
fLanguage
English
Publisher
ieee
Conference_Titel
Machine Learning for Signal Processing, 2005 IEEE Workshop on
Conference_Location
Mystic, CT
Print_ISBN
0-7803-9517-4
Type
conf
DOI
10.1109/MLSP.2005.1532890
Filename
1532890
Link To Document