论文部分内容阅读
网络是表达物体和物体间联系的一种重要形式,针对网络的分析研究的一个关键问题就是研究如何合理地表示网络中的特征信息.随着机器学习技术的发展,针对网络中节点的特征学习成为了一项新兴的研究任务.网络表示学习算法将网络信息转化为低维稠密的实数向量,并用于已有的机器学习算法的输入.举例来说,节点表示可以作为特征送入支持向量机等分类器用于节点分类任务,也可以作为欧氏空间中的点坐标用于可视化任务.近年来,网络表示学习问题吸引了大量的研究者的目光,本文将针对近年来的网络表示学习工作进行系统性的介绍和总结.
Network is an important form to express the connection between objects and objects. A key problem for network analysis and study is to study how to represent characteristic information in network reasonably. With the development of machine learning technology, Has become an emerging research task.The network indicates that the learning algorithm transforms the network information into a low-dimensional dense real vector and is used for the input of existing machine learning algorithms.For example, the node representation can be used as a feature into the SVM The classifier is used for the task of node classification and also can be used as the point coordinates in Euclidean space for visualization tasks.In recent years, the network representation learning problem has attracted a great deal of researchers’ eyes. This paper will focus on the recent network representation learning work Systematic introduction and summary.