In the smart cities of the future artificial intelligence (AI) will have a dominant role given that AI will accommodate the utilization of intelligent analytics for prediction of critical parameters pertaining to city operation. In this chapter, a new data analytics paradigm is presented and being applied for energy demand forecasting in smart cities. In particular, the presented paradigm integrates a group of kernel machines by utilizing a deep architecture. The goal of the deep architecture is to exploit the strong capabilities of deep learning utilizing various abstraction levels and subsequently identify patterns of interest in the data. In particular, a deep feedforward deep neural network is employed with every network node to implement a kernel machine. This deep architecture, named neuro-kernel machine network, is subsequently applied for predicting the energy consumption of groups of residents in smart cities. Obtained results exhibit the capability of the presented method to provide adequately accurate predictions despite the form of the energy consumption data.
CITATION STYLE
Alamaniotis, M. (2021). Neuro-kernel-machine network utilizing deep learning and its application in predictive analytics in smart city energy consumption. In Intelligent Systems Reference Library (Vol. 189, pp. 293–307). Springer. https://doi.org/10.1007/978-3-030-51870-7_14
Mendeley helps you to discover research relevant for your work.