Urban air pollutant concentration prediction is dealing with a surge of massive environmental monitoring data and complex changes in air pollutants. This requires effective prediction methods to improve prediction accuracy and to prevent serious pollution incidents, thereby enhancing environmental management decision-making capacity. In this paper, a new pollutant concentration prediction method is proposed based on the vast amounts of environmental data and deep learning techniques. The proposed method integrates big data by using two kinds of deep networks. This method is based on the design that uses a convolutional neural network as the base layer, automatically extracting features of input data. A long shortterm memory network is used for the output layer to consider the time dependence of pollutants. Our model consists of these two deep networks. With performance optimization, the model can predict future particulate matter (PM 2.5) concentrations as a time series. Finally, the prediction results are compared with the results of numerical models. The applicability and advantages of the model are also analyzed. The experimental results show that it improves prediction performance compared with classic models. INDEX TERMS Air pollution, machine learning, neural network, numerical analysis, prediction method.
For a (molecular) graph, the first Zagreb index M 1 is equal to the sum of squares of its vertex degrees, and the second Zagreb index M 2 is equal to the sum of products of degrees of pairs of adjacent vertices. A connected graph G is a cactus if any two of its cycles have at most one common vertex. In this paper, we investigate the first and the second Zagreb indices of cacti with k pendant vertices. We determine sharp bounds for M 1-, M 2-values of n-vertex cacti with k pendant vertices. As a consequence, we determine the n-vertex cacti with maximal Zagreb indices and we also determine the cactus with a perfect matching having maximal Zagreb indices.
With the development of cloud computing, the cloud market is becoming more and more complicated. In a cloud data center, there are many cloud instance types with different computing capacity and price, which brings users some confusions when they select cloud instance types. In order to solve this selection problem, a cloud brokering architecture is proposed. In this architecture, the selection problem is modeled as a multi-objective optimization problem, and through analysis, we get the relationship between complete Pareto set and solution space. Based on this, a two-stage Cloud Instance Type Selection Model (CITSM) is proposed to help users select the cloud instance types. The first stage is Complete Pareto Set Generation Algorithm (CPSGA) which can generate a complete Pareto set of the cloud instance type selection schemes. Then, the Optimal cloud instance type selection Scheme Screening Algorithm (OSSA) is used to select one scheme from the complete Pareto set. We perform some experiments to prove the proposed CITSM is efficient and effective. The proposed method can also solve the single objective optimization problem by modifying OSSA, which illustrates the scalability.INDEX TERMS Cloud computing, instance type selection, cloud broker, multi-objective optimization, complete Pareto set.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.