Regional international cooperation in seismic data exchange and joint seismological analysis in the Asian-Pacific regions plays an important role in the reduction of earthquake disasters in these regions as well as th...Regional international cooperation in seismic data exchange and joint seismological analysis in the Asian-Pacific regions plays an important role in the reduction of earthquake disasters in these regions as well as the development of global seismology. Based on the up-to-date development of new technology and digital broadband seismology, we discussed the Asian Seismological Commission (ASC) Proposal “Asian-Pacific Seismological Data Center”. We applied the concept of virtual seismological network (Ottemoeller and Havskov, 1999) to the proposed ASC data exchange program. Based on the development of digital seismology, we introduced a concept of “modern earthquake catalogues” which include not only location parameters and magnitudes as in the conventional earthquake catalogues but also new parameters of earthquakes such as CMT, radiated energy, STF, and earthquake rupture process. We recommended a web-based regional international data exchange program between the ASC members with the purpose of forming a virtual Asian-Pacific seismological network, and the interpretation and analysis of broadband digital seismic waveforms conducted at distributed “data centers”.展开更多
随着基于位置的社交网络的快速发展,下一个PoI(point of interest)推荐已成为推荐领域的研究热点。然而现有研究模型忽略了PoI的时空特征以及上下文信息对下一个PoI推荐的效果。针对该问题,提出一种时空上下文感知的下一个PoI推荐方法...随着基于位置的社交网络的快速发展,下一个PoI(point of interest)推荐已成为推荐领域的研究热点。然而现有研究模型忽略了PoI的时空特征以及上下文信息对下一个PoI推荐的效果。针对该问题,提出一种时空上下文感知的下一个PoI推荐方法。首先,利用图注意力网络(GAT)学习包含社交关系的用户表征;并且通过流行度增强二部图神经网络(PEBGNN)学习含有PoI交互偏好的用户表征和PoI表征;同时,利用时空图卷积网络(ST-GCN)学习PoI时空转移偏好的PoI表征;最后,通过融合所学到的用户表征和PoI表征,计算出用户对于各个PoI的预测评分,以此为基础为用户推荐下一个PoI。为了验证该方法的有效性,在Gowalla、Foursquare以及Yelp这三个公开的数据集上进行了测试。实验结果显示,相比于多个基准模型,所提方法在准确率和召回率方面均展现出了显著的优势,分别平均提升28.53%和7.65%。展开更多
文摘Regional international cooperation in seismic data exchange and joint seismological analysis in the Asian-Pacific regions plays an important role in the reduction of earthquake disasters in these regions as well as the development of global seismology. Based on the up-to-date development of new technology and digital broadband seismology, we discussed the Asian Seismological Commission (ASC) Proposal “Asian-Pacific Seismological Data Center”. We applied the concept of virtual seismological network (Ottemoeller and Havskov, 1999) to the proposed ASC data exchange program. Based on the development of digital seismology, we introduced a concept of “modern earthquake catalogues” which include not only location parameters and magnitudes as in the conventional earthquake catalogues but also new parameters of earthquakes such as CMT, radiated energy, STF, and earthquake rupture process. We recommended a web-based regional international data exchange program between the ASC members with the purpose of forming a virtual Asian-Pacific seismological network, and the interpretation and analysis of broadband digital seismic waveforms conducted at distributed “data centers”.
文摘随着基于位置的社交网络的快速发展,下一个PoI(point of interest)推荐已成为推荐领域的研究热点。然而现有研究模型忽略了PoI的时空特征以及上下文信息对下一个PoI推荐的效果。针对该问题,提出一种时空上下文感知的下一个PoI推荐方法。首先,利用图注意力网络(GAT)学习包含社交关系的用户表征;并且通过流行度增强二部图神经网络(PEBGNN)学习含有PoI交互偏好的用户表征和PoI表征;同时,利用时空图卷积网络(ST-GCN)学习PoI时空转移偏好的PoI表征;最后,通过融合所学到的用户表征和PoI表征,计算出用户对于各个PoI的预测评分,以此为基础为用户推荐下一个PoI。为了验证该方法的有效性,在Gowalla、Foursquare以及Yelp这三个公开的数据集上进行了测试。实验结果显示,相比于多个基准模型,所提方法在准确率和召回率方面均展现出了显著的优势,分别平均提升28.53%和7.65%。
