华北地震电离层观测网的沈阳周边地区特性分析

为更好掌握华北地震观测网所得到的空间环境特性,为地震前兆与电离层特性的相关性分析提供依据。以沈阳地震台电离层斜测站接收到长春-沈阳、北京-沈阳、新乡-沈阳、苏州-沈阳等探测链路的最高可用频率(MUF)为基础,利用电离层反演方法获取路径中心的电离层临界频率,并通过非线性时间序列分析方法和多维空间重构方法,分析了与地震相关的电离层F_2临界频率的时空变化规律:a.6时左右出现最小值,逐渐增大直至正午前后达到最大值,再逐渐减小至第二天日出。b.冬季的昼夜极差最大,夏季的昼夜极差最小。链路距离较长的昼夜极差相对较大,链路距离较短的昼夜极差相对较小。c.冬季幅度值较高,春秋季幅度值较小。d.纬度较高的随季节变化明显,纬度较低的随季节变化较为平缓。上述研究可为地震前期电离层异常特征分析及其与地震前兆研究提供基础。     

基于GEP的经济时间序列组合预测方法研究

线性组合预测效果欠佳,非线性函数的挖掘也很困难,本文提出了基于基因表达式编程的非线性组合预测的新方法.理论分析和应用实例表明,相比模糊神经网络等组合预测而言,该方法具有很强的学习和仿真功能,在社会经济复杂系统中时间序列的组合建模和预测中具有很好的应用价值.     

基于图卷积神经网络的图书推荐方法研究

[目的/意义]用户与项目的交互历史中包含大量的语义信息,现有的协同过滤方法无法捕获其中的信息,将具有良好特征表示能力的图卷积神经网络引入图书推荐领域。[方法/过程]根据交互历史构建读者-图书二部图,搭建图卷积神经网络,通过连续的卷积层捕获二部图的高阶连通性来得到读者的邻域偏好信息,在预测层对邻域信息聚合并开展预测。[结果/结论]将提出的方法与对比方法在豆瓣图书数据集上进行实验,结果表明所提出的基于图卷积神经网络的图书推荐方法在召回率和归一化折损累计增益两项指标上均取得更好的表现。     

电离层探测系统数据管理与显示软件的设计与开发

电离层垂直探测是地基电离层探测研究中获取电离层参数最直接的方法,因此成为研究电离层的重要手段。为了观测研究我国西南地区电离层,中国地震局预测研究所,中国科学院空间中心等单位联合武汉大学电离层实验室在云南普洱,四川乐山,海南三亚和甘肃张掖等地建立电离层观测站点。为了有效管理电离层垂测站记录的频高图数据,并实时监测各观测站点的基本工作状态,本文设计开发了电离层探测系统数据管理与显示软件。该软件通过观测站记录的电离层foF2参数,对国际参考电离层模型(International Reference Ionosphere,IRI)~([1])进行修正,实时提供我国西南地区的foF2态势变化图。同时,该软件通过显示各观测站点的频高图,来远程监测各个站点垂测仪设备的基本工作状态,以方便探测设备和数据的日常管理与维护。     

基于神经网络的高校艾滋病发病率预测分析

利用传统算法进行艾滋病发病率预测分析过程中,由于传统数据的处理方法具有一定的模糊性,一定程度上降低了高校艾滋病发病率预测分析的准确性。提出基于神经网络的高校艾滋病发病率预测分析优化算法。根据非线性函数的原理实现艾滋病发病率的输入数据从神经网络的输入层到输出层的快速传递,通过合理地调节各个神经层中艾滋病发病率的神经元之间的权系数,能够有效的实现高校艾滋病发病率的预测分析。实验结果表明,利用改进算法进行高校艾滋病发病率的预测分析,极大程度上提高了预测分析系统中输入数据与输出数据的分析精度。     

神经网络在足球比赛中的胜负预测

因人工神经网络具有极强的非线性逼近能力,所以在诸多领域中得到了广泛的应用。足球作为世界第一大运动,相关专家及球迷一直热衷于对其比赛的结果进行预测。本文提出了一种使用神经网络来预测比赛胜负的新方法,并用R语言进行实现并在相关数据集上进行测试,通过测试发现所提出的新预测方法具有一定的可靠性。     

房地产销售价格指数的神经网络预测

房地产销售价格指数是指导业界活动和市场研究的有效工具,但是预测的准确程度一直是人们倍加关注的。人工神经网络是一门新兴交叉学科,近年来被越来越多的应用到了实际问题的预测中,显示出其广阔的应用前景,特别是人工神经网络具有预测非线性系统未来行为的巨大潜力。因此,本文提出了用人工神经网络对房地产销售价格指数进行预测的方法,首先将输入数据进行预处理,再利用多层前馈神经网络BP算法来研究人工神经网络在房地产销售价格指数预测中的应用问题,最后得出神经网络方法预测精度较高的结论。     

日本电离层频高图转换方法研究

为了从日本电离层频高图中准确地获取特征参数和电子密度剖面,提出了利用数字图像处理等技术对频高图进行转换的方法。该方法是利用Radon变换、最大类间方差法和连通分量标记法等对频高图进行滤噪优化和格式转换。利用该方法开发出频高图批量处理软件,并结合SAO Explorer软件对日本Okinawa地区的频高图进行电离层特征参数的提取和电子密度剖面的自动度量实例分析,通过对转换前后特征参数f0F2的统计分析得到可接受率分别为52.1%和90.1%。结果表明,该方法具有较高的可靠性和精确性。     

基于综合测量的岩土坡体稳定裕度的研究

针对目前现有的地质滑坡的形成条件、诱发因素错综复杂,使用传统的测量手段存在实时性差,准确度低的情况,提出一种基于BP神经网络的滑坡预警模型。通过多个节点传感器综合测量得到滑坡发生的输入参数,对多组输入参数进行降维处理,并使用遗传算法调整输入层、隐含层以及输出层的权值和阈值,提高神经网络的训练精度,使得输出层的预测值更接近理想期望值。通过对多组神经网络的样本训练之后,对训练后的神经网络进行泛化,并将其投入到实际的应用场景,对地质滑坡进行预测分析。同时,搭建滑坡预警的信息采集显示系统,通过下位机多个节点采集传感器信号,传至主节点加以显示并在上位机监测软件实时显示。通过实验验证采用神经网络的智能学习算法,得到的预测结果与实际情况基本一致,验证了神经网络对于滑坡预测的切实可行性。     

甚低频电磁波在地-电离层波导中的场强预测

甚低频电波传播特性分析与预测是提高甚低频导航/授时系统精度的关键。本文基于甚低频电磁波在"地-电离层"波导中的传播理论,结合IRI模型和NRLMSISE-00大气模型以及传播矩阵方法,构建了甚低频电波传播场强实时预测模型。对比了不同电离层模型间的差异以及电离层参数的时变特性,仿真预测了接收点场强及其日变化情况,并与文献实测结果进行了比较。计算结果表明:本文预测结果与实测数据吻合较好,较采用电离层指数模型的传统解析方法精度得到明显提升,可更有效地反映接收点场强的时变特性。     

Boundaries and edges rethinking: An end-to-end neural model for overlapping entity relation extraction

Overlapping entity relation extraction has received extensive research attention in recent years. However, existing methods suffer from the limitation of long-distance dependencies between entities, and fail to extract the relations when the overlapping situation is relatively complex. This issue limits the performance of the task. In this paper, we propose an end-to-end neural model for overlapping relation extraction by treating the task as a quintuple prediction problem. The proposed method first constructs the entity graphs by enumerating possible candidate spans, then models the relational graphs between entities via a graph attention model. Experimental results on five benchmark datasets show that the proposed model achieves the current best performance, outperforming previous methods and baseline systems by a large margin. Further analysis shows that our model can effectively capture the long-distance dependencies between entities in a long sentence.     

基于蒙特卡洛模拟和挣值分析的项目完工预测优化

基于挣值分析和风险管理，通过蒙特卡洛模拟获取项目数据，使用二次判别分析、随机森林和支持向量机进行模型学习和完工预测是项目控制的有效方法之一。在现有研究基础上，考虑项目执行过程中的剩余工作时间、剩余工作费用和风险，分别应用现有研究方法、梯度提升树和人工神经网络进行模型学习，利用嵌套交叉验证进行模型选择和模型评估。研究结果表明，优化后的方法显著提升项目完工预测的准确率。     

基于专利文献和知识图谱的技术预测方法研究

在分析现有基于专利文献进行技术预测方法不足的基础上,提出一种基于专利文献和知识图谱的技术预测方法。(1)使用Google知识图谱和领域知识创建领域知识图谱;(2)依据创建的领域知识图谱对专利文献赋予标签;(3)引入社会网络社区进化研究成果,基于专利文献标签之间的网络图进行新兴技术预测。以肺癌领域技术预测为例,绘制肺癌领域知识图谱,进行方法验证并预测。验证结果显示,该方法可较好地进行技术预测。     

The role of location and social strength for friendship prediction in location-based social networks

Recent advances in data mining and machine learning techniques are focused on exploiting location data. These advances, combined with the increased availability of location-acquisition technology, have encouraged social networking services to offer to their users different ways to share their location information. These social networks, called location-based social networks (LBSNs), have attracted millions of users and the attention of the research community. One fundamental task in the LBSN context is the friendship prediction due to its role in different applications such as recommendation systems. In the literature exists a variety of friendship prediction methods for LBSNs, but most of them give more importance to the location information of users and disregard the strength of relationships existing between these users. The contributions of this article are threefold, we: 1) carried out a comprehensive survey of methods for friendship prediction in LBSNs and proposed a taxonomy to organize the existing methods; 2) put forward a proposal of five new methods addressing gaps identified in our survey while striving to find a balance between optimizing computational resources and improving the predictive power; and 3) used a comprehensive evaluation to quantify the prediction abilities of ten current methods and our five proposals and selected the top-5 friendship prediction methods for LBSNs. We thus present a general panorama of friendship prediction task in the LBSN domain with balanced depth so as to facilitate research and real-world application design regarding this important issue.     

基于最优权值探测的图像可逆信息隐藏研究

提出一种基于最优权值探测的图像可逆信息隐藏算法。该算法改进了现有的预测方法,利用最优权值探测原理提高预测像素值的精确受,使预测误差差值直方图更加紧凑、峰值更高。实验结果表明,相比其它可逆嵌入算法．该算法在保证图像质量的同时,提高了嵌入容量。算法整体性能高。     

A new method for fault prediction of model-unknown nonlinear system

In this paper, the subspace identification based robust fault prediction method which combines optimal track control with adaptive neural network compensation is presented for prediction the fault of unknown nonlinear system. At first, the local approximate linear model based on input-output of unknown system is obtained by subspace identification. The optimal track control is adopted for the approximate model with some unknown uncertainties and external disturbances. An adaptive RBF neural network is added to the track control in order to guarantee the robust tracking ability of the observation system. The effect of the system nonlinearity and the error caused by subspace modeling can be overcome by adaptive tuning of the weights of the RBF neural network online without any requisition of constraint or matching conditions. The stability of the designed closed-loop system is thus proved. A density function estimation method based on state forecasting is then used to judge the fault. The proposed method is applied to fault prediction of model-unknown fighter F-8II of China airforce and the simulation results show that the proposed method can not only predict the fault, but has strong robustness against uncertainties and external disturbances.     

Predictor-based optimal robust guaranteed cost control for uncertain nonlinear systems with completely tracking errors constraint

In this paper, a novel composite controller is proposed to achieve the prescribed performance of completely tracking errors for a class of uncertain nonlinear systems. The proposed controller contains a feedforward controller and a feedback controller. The feedforward controller is constructed by incorporating the prescribed performance function (PPF) and a state predictor into the neural dynamic surface approach to guarantee the transient and steady-state responses of completely tracking errors within prescribed boundaries. Different from the traditional adaptive laws which are commonly updated by the system tracking error, the state predictor uses the prediction error to update the neural network (NN) weights such that a smooth and fast approximation for the unknown nonlinearity can be obtained without incurring high-frequency oscillations. Since the uncertainties existing in the system may influence the prescribed performance of tracking error and the estimation accuracy of NN, an optimal robust guaranteed cost control (ORGCC) is designed as the feedback controller to make the closed-loop system robustly stable and further guarantee that the system cost function is not more than a specified upper bound. The stabilities of the whole closed-loop control system is certified by the Lyapunov theory. Simulation and experimental results based on a servomechanism are conducted to demonstrate the effectiveness of the proposed method.     

神经网络和预报集成的若干问题及其对策

本文评述了神经网络在气象应用中存在的问题、产生的原因和网络结构设计的原则及经验,并提出了构造伪样本来诊断模型可能存在问题的方法;简述了预报集成的原理和原则,提出了构造因子子集差异法生成预报个体,进行集成,来解决小样本和因子选取问题的方案。     

Multi-Sentence Video Captioning using Content-oriented Beam Searching and Multi-stage Refining Algorithm

With the increasing growth of video data, especially in cyberspace, video captioning or the representation of video data in the form of natural language has been receiving an increasing amount of interest in several applications like video retrieval, action recognition, and video understanding, to name a few. In recent years, deep neural networks have been successfully applied for the task of video captioning. However, most existing methods describe a video clip using only one sentence that may not correctly cover the semantic content of the video clip. In this paper, a new multi-sentence video captioning algorithm is proposed using a content-oriented beam search approach and a multi-stage refining method. We use a new content-oriented beam search algorithm to update the probabilities of words generated by the trained deep networks. The proposed beam search algorithm leverages the high-level semantic information of an input video using an object detector and the structural dictionary of sentences. We also use a multi-stage refining approach to remove structurally wrong sentences as well as sentences that are less related to the semantic content of the video. To this intent, a new two-branch deep neural network is proposed to measure the relevance score between a sentence and a video. We evaluated the performance of the proposed method with two popular video captioning databases and compared the results with the results of some state-of-the-art approaches. The experiments showed the superior performance of the proposed algorithm. For instance, in the MSVD database, the proposed method shows an enhancement of 6% for the best-1 sentences in comparison to the best state-of-the-art alternative.     

Aspect sentiment analysis with heterogeneous graph neural networks

Aspect-based sentiment analysis technologies may be a very practical methodology for securities trading, commodity sales, movie rating websites, etc. Most recent studies adopt the recurrent neural network or attention-based neural network methods to infer aspect sentiment using opinion context terms and sentence dependency trees. However, due to a sentence often having multiple aspects sentiment representation, these models are hard to achieve satisfactory classification results. In this paper, we discuss these problems by encoding sentence syntax tree, words relations and opinion dictionary information in a unified framework. We called this method heterogeneous graph neural networks (Hete_GNNs). Firstly, we adopt the interactive aspect words and contexts to encode the sentence sequence information for parameter sharing. Then, we utilized a novel heterogeneous graph neural network for encoding these sentences’ syntax dependency tree, prior sentiment dictionary, and some part-of-speech tagging information for sentiment prediction. We perform the Hete_GNNs sentiment judgment and report the experiments on five domain datasets, and the results confirm that the heterogeneous context information can be better captured with heterogeneous graph neural networks. The improvement of the proposed method is demonstrated by aspect sentiment classification task comparison.