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 prevent serious pollution incidents, thereby enhancing environmental management decision-making capacity. In this paper, a new pollutant concentration prediction method is proposed based on vast amounts of environmental data and deep learning techniques. The proposed method integrates big data using two kinds of deep networks. This method is based on a design that uses a Convolutional Neural Network as the base layer, automatically extracting features of input data. A Long Short Term 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 (PM2:5) concentrations as time series. Finally, the prediction results are compared with the results of numerical models. The applicability and advantages of the model are also analyzed. Experimental results show that it improves prediction performance compared with classic models.
Most community detection methods focus on the similarities between detection nodes to achieve community partitioning. Traditional network representation learning methods are also limited to the local context of the central nodes, which results in less truly representative results. This article examines nodes' influence information, nodes' community affiliating information, and similarity of community topologies and proposes a more effective node representation strategy. According to the local node information and global topology in the social network graph, a method of combining local node embedding and global community embedding is also designed. The effectiveness of learning node representation and community representation is improved by our approach. The proposed model can also effectively detect overlapping communities.
26 August 2024 in Clinical and Translational Medicine
Location-based recommendation has recently received a lot of attention in the communities of information service and mobile application. Its task is to provide personalized recommendations of points of interest (POIs) to users at a certain time and location. However, existing location-based recommendation models have at least two main drawbacks: first they cannot adequately capture semantic features of POIs and users, which may lead to unsatisfactory recommendations and second they cannot effectively address the cold-start problem. To address the above drawbacks, in this article, we first propose a novel deep representation learning-based model (DRLM) for improving the recommendation accuracy. In DRLM, we mainly focus on learning to accurately represent semantic features of POIs and users. Specifically, four co-occurrence matrices are constructed to produce four different original features for each POI, and a principal component analysis (PCA) algorithm is utilized to generate a semantic feature of each POI from its four original features. On the other hand, a three-modal simple recurrent unit (TMSRU) network is given to constructed semantic features of users using semantic features of POIs, times, and locations. We further propose minimum description length (MDL)-based and skyline-based strategies to address the cold-start issues for new users and new POIs, respectively. Through experiments on two real-world data sets, we show that compared with the state-of-the-art approaches, the proposed model DRLM can achieve the superior performance in terms of high recommendation accuracy and effectiveness in handling the cold-start problem.