Wu, Chao-LinChao-LinWuChen, Wei-ChenWei-ChenChenTseng, Yi-ShowYi-ShowTsengFu, Li-ChenLi-ChenFuLI-CHEN FU2020-05-042020-05-042014https://scholars.lib.ntu.edu.tw/handle/123456789/488993A traditional context-aware energy-saving system is often re-active, which means the decision of the system is made purely based on the currently available contexts of the users in the sensed environment. However, the advances in internet-of-things (IoTs) enable the potentials of leveraging predictive contexts to facilitate proactive energy-savings services. Therefore, this paper proposed hybrid context prediction by using previously learned periodical and sequential patterns to achieve anticipatory reasoning. The evaluation results show that the proposed approach will lead to more acceptable pro-active energy-saving services as well as more desirable user comfort. © 2014 IEEE.[SDGs]SDG7Internet; Internet of things; Activity predictions; Anticipatory reasoning; Context predictions; Energy saving systems; Evaluation results; Internet of thing (IoTs); Predictive context; Sequential patterns; Energy conservationconference paper10.1109/iThings.2014.41https://doi.org/10.1109/iThings.2014.41