The amount of information increases explosively in Internet of Things, because more and more data are sensed by large amount of sensors. The explosive growth of information makes it difficult to access information efficiently, so it is an effective method to decrease the amount of information to be transferred on network by text classification. This paper proposes a new text classification algorithm based on vector space model. This algorithm improves the feature selection and weighting methods by introducing synonym replacement to traditional text classification algorithms. The experimental results show that the proposed classification algorithm has considerably improved the precision and recall of classification.

No abstract

No abstract

With pumps as the main devices, the main drainage system (MDS) is critical to mine construction and production. Considering the high cost of traditional manual pump scheduling strategy and the wide adoption of time-of-use (TOU) electricity traffic in coal mines, this paper attempts to reduce the mine operation cost by scheduling the pumps in the flat and valley periods instead of the peak period. For this purpose, the pump scheduling was considered as an optimization problem, the water level of the sump was predicted by double exponential smoothing, and then the optimal pump scheduling plan was derived by ant colony optimization (ACO). The pump scheduling plan obtained by the proposed method was proved cost efficient through experiments on a gold mine in China.