Comparative study of multilayer perceptron-stochastic gradient descent and gradient boosted trees for predicting daily suspended sediment load: The case study of the Mississippi River,U.S.

Monitoring sediment transport is essential for managing and maintaining rivers.Estimation of the sediment load in rivers is fundamental for the study of sediment movement,erosion,and flood control.In the current study,three machine learning models-multi-layer perceptron(MLP),multi-layer perceptron-stochastic gradient descent(MLP-SGD),and gradient boosted tree(GBT)-were utilized to estimate the suspended sediment load(SSL)at the St.Louis(SL)and Chester(CH)stations on the Mississippi River,U.S.Four evaluation criteria including the Correlation Coefficient(CC),Nash Sutcliffe Efficiency(NSE),Scatter Index(SI),and Willmott’s Index(WI)were utilized to evaluate the performance of the used models.A sensitivity analysis of the models to the input variables revealed that the current day discharge variable had the most effect on the SSL at both stations,but in the absence of current-day discharge data(Qt),a combination of input parameters including SSL_t-3,SSL_t-2,SSL_t-1,Q_t-3,Q_t-2,Q_t-1 can be used to estimate the SSL.The comparative outcomes indicated the high accuracy of MLP-SGD-5 model with a CC of 0.983,SI of 0.254,WI of 0.991,and NSE of 0.967 at station CH and the MLP-SGD-6 model with a CC of 0.933,SI of 0.576,WI of 0.961,and NSE of 0.867,respectively,at station SL.The results of MLP models were improved by SGD optimization.Therefore,the MLP-SGD method is recommended as the most accurate model for SSL estimation.     

ST‐DMQL: A Semantic Trajectory Data Mining Query Language

Mobile devices are becoming very popular in recent years, and large amounts of trajectory data are generated by these devices. Trajectories left behind cars, humans, birds or other objects are a new kind of data which can be very useful in the decision making process in several application domains. These data, however, are normally available as sample points, and therefore have very little or no semantics. The analysis and knowledge extraction from trajectory sample points is very difficult from the user's point of view, and there is an emerging need for new data models, manipulation techniques, and tools to extract meaningful patterns from these data. In this paper we propose a new methodology for knowledge discovery from trajectories. We propose through a semantic trajectory data mining query language several functionalities to select, preprocess, and transform trajectory sample points into semantic trajectories at higher abstraction levels, in order to allow the user to extract meaningful, understandable, and useful patterns from trajectories. We claim that meaningful patterns can only be extracted from trajectories if the background geographical information is considered. Therefore we build the proposed methodology considering both moving object data and geographic information. The proposed language has been implemented in a toolkit in order to provide a first software prototype for trajectory knowledge discovery.     

A method for discovery and analysis of temporal patterns in complex event data

Pattern analysis techniques currently common within geography tend to focus either on characterizing patterns of spatial and/or temporal recurrence of a single event type (e.g., incidence of flu cases) or on comparing sequences of a limited number of event types where relationships between events are already represented in the data (e.g., movement patterns). The availability of large amounts of multivariate spatiotemporal data, however, requires new methods for pattern analysis. Here, we present a technique for finding associations among many different event types where the associations among these varying event types are not explicitly represented in the data or known in advance. This pattern discovery method, known as T-pattern analysis, was first developed within the field of psychology for the purpose of finding patterns in personal interactions. We have adapted and extended the T-pattern method to take the unique characteristics of geographic data into account and implemented it within a geovisualization toolkit for an integrated computational-geovisual environment we call STempo. To demonstrate how T-pattern analysis can be employed in geographic research for discovering patterns in complex spatiotemporal data, we describe a case study featuring events from news reports about Yemen during the Arab Spring of 2011–2012. Using supplementary data from the Global Database of Events, Language, and Tone, we briefly summarize and reference a separate validation study, then evaluate the scalability of the T-pattern approach. We conclude with ideas for further extensions of the T-pattern technique to increase its utility for spatiotemporal analysis.     

基于多任务学习的测井储层参数预测方法

基于多任务神经网络模型, 提出一种多任务测井储层参数预测方法, 利用测井数据对储层孔隙度、渗透率及含水饱和度同时进行预测.分别采用同架构和异架构多任务模型对测井储层参数进行预测, 通过数值实验对比, 多任务预测模型有效提升了单任务储层参数预测模型的效果, 且提升幅度与模型结构有关, 异架构多任务模型的总体预测效果好于同架构多任务模型.以平均相对误差(MAPE)作为模型评价标准, 针对本研究所采用的数据集, 同架构多任务模型的孔隙度、渗透率和含水饱和度在测试集上的MAPE约为6%、17%和9%, 相较于单任务模型, 预测效果分别提升约30%、20%和10%.异架构多任务模型的孔隙度、渗透率和含水饱和度, 在测试集上的MAPE约为6%、13%和6%, 相较于单任务模型分别提升超过2%、60%和10%.

     

一种新颖的探地雷达快速正演模拟及埋地目标探测机器学习方法

探地雷达正演模拟在真实雷达数据解译及全波形反演中扮演着重要的角色,针对传统探地雷达（Ground Penetrating Radar,GPR）正演模拟计算量巨大、耗时、不利于实时探测等问题,提出一种基于机器学习框架的近实时GPR正演模拟方法.以混凝土中的钢筋探测作为GPR应用场景,混凝土的含水量、钢筋半径及埋地深度作为模型参数,利用时域有限差分数值模拟散射回波信号;运用主成分分析对回波数据进行降维处理得到相应的主成分权值系数,并以此作为机器学习网络的输出;设计了一种基于随机森林的多层循环网络架构和学习策略,不仅充分挖掘学习模型参数和主成分权值系数之间的内在因果关系,也共享主成分间的相互联系,并具有对每个预测主成分完善和修正的功能,以此实现基于机器学习的探地雷达快速正演模拟,与传统机器学习相比,有效提高了正演模拟的精度.在此基础上将两个深度神经网络与随机森林相结合,以回波数据主成分系数为输入,建立了基于机器学习的场景参数预测模型,实现了近实时的埋地目标探测,预测的混凝土含水量最大误差为2%,钢筋埋地深度最大误差为6.7%.

     

地震序列类型的确定与现场预报规则的获取

论述了“震后趋势决策支持系统PTDSS”中的知识学习问题。以某些地震参数与序列类型的关系为例,介绍了如何确定地震序列类型以及系统通过FAM模型进行机器学习的方法。通过学习系统得到了一批非常有用的早期判断序列类型的知识。     

地球科学的新使命——认知和发现非传统矿产资源

非传统矿产资源是指受目前经济、技术以及环境因素的限制尚未发现和尚未开发利用的矿产资源及尚未被看作矿产，未发现其用途的潜在矿产资源，面对矿产资源需求供给日益严峻的局面，认知、发现和开发利用非传统矿产资源是一个十分重要、但目前尚不被广泛重视、也未开展系统规划和研究的领域。作者指出：（1）“新的地球物质资源观”，即地球的一切固、液、气体物质均为资源，只有已认知、已发现及已利用资源和尚未被认知、尚未被发现和尚未被利用资源之分，资源也不是一个绝对的概念，甚至资源已不仅局限于地球物质，（2）地球物质资源是随时间、空间而变化的技术、经济、环境、市场、人文等因素的函数，资源是一个随时间推移和空间不同而变化的动态概念，“传统”和“非传统”资源也是随时间而变化的概念；（3）非传统矿产资源研究与开发是一项具有前瞻性和战略性的工作，是一种更长远的资源战略储备；近期研究目标主要包括非传统矿产资源发现和矿产资源非传统方法研究两个方面。     

EMD在广西季节降水预报中的应用

气候系统是一种耗散的、具有多个不稳定源的非线性、非平稳系统。该文利用支持向量机(SVM)算法在处理非线性问题中的优越性和经验模态分解(EMD)算法在处理非平稳信号中的优势,采用将EMD与SVM相结合的短期气候预测方法,并应用到广西季节降水预报中。选取广西88个气象观测站1957—2005年6—8月逐年降水量的距平百分率序列作为试验数据,通过EMD算法将标准化处理后的距平百分率序列分解成多个本征模态函数(IMF)分量和一个趋势分量,在分解中针对EMD算法存在的端点极值问题选择两种方法分别进行处理,对比得出极值延拓法效果更好。对每个分量构建不同的SVM模型进行预测,并通过重构形成最后的预测结果。试验中采用不经EMD处理的反向传播(BP)神经网络和SVM算法进行对比验证,结果表明:相对于直接预测方法,该文提出的方案均方误差最小,能够较为准确地反映出降水序列未来几年的变化趋势,具有更高的预测精度和较好的推广前景。     

处理非线性分类和回归问题的一种新方法（Ⅰ）——支持向量机方法简介

简要介绍了近年来倍受瞩目的一种处理高度非线性分类、回归等问题的计算机学习的新方法——支持向量机(SVM)方法；分析了这一方法的特点及其在数值预报产品释用及气象研究业务中的应用前景。SVM是一种有坚实理论基础的新颖的小样本学习方法。它基本上不涉及概率测度及大数定律等，因此不同于现有的统计方法。从本质上看，它避开了从归纳到演绎的传统过程，实现了高效的从训练样本到预报样本的“转导推理”(transductive inference)，大大简化了通常的分类和回归等问题。SVM的最终决策函数只由少数的支持向量所确定，计算的复杂性取决于支持向量的数目，而不是样本空间的维数，这在某种意义上避免了“维数灾”。     

Machine Learning of Weather Forecasting Rules from Large Meteorological Data Bases

Discovery of useful forecasting rules from observational weather data is an outstanding interesting topic.The traditional methods of acquiring forecasting knowledge are manual analysis and investigation performed by human scientists.This paper presents the experimental results of an automatic machine learning system which derives forecasting rules from real observational data.We tested the system on the two large real data sets from the areas of centra! China and Victoria of Australia.The experimental results show that the forecasting rules discovered by the system are very competitive to human experts.The forecasting accuracy rates are 86.4% and 78% of the two data sets respectively