An object-oriented knowledge representation structure for exploration data integration

Knowledge representation structure and reasoning processes are very important issues in the knowledge-based approach of integrating multiple spatial data sets for resource exploration. An object-oriented knowledge representation structure and corresponding reasoning processes are formulated and tested in this research on the knowledge-based approach of integrating spatial exploration data. The map-based prototype expert system developed in this study has self-contained knowledge representation structure and inference mechanisms. It is important to distinguish between lack of information and information providing negative evidence for a map-based system because the spatial distribution of data sets are uneven in most cases. Error and uncertainty estimation is also an important component of any production expert system. The uncertainty propagation mechanisms developed here work well for this type of integrated exploration problem. Evidential bellef function theory provides a natural theoretical basis for representing and integrating spatially uneven geophysical and geological information. The prototype system is tested using real mineral exploration data sets from the Snow Lake area, northern Manitoba, Canada. The test results outline the favorable exploration areas successfully and show the effectiveness of the knowledge representation structure and inference mechanisms for the knowledge-based approach.     

基于知识表示的DDES辅助决策系统的结构设计

介绍了土地合理利用规划决策专家系统(DDES)辅助决策系统的构建原理与实现方法,讨论了DDES中专家知识的表示方法与推理控制策略。系统将领域专家知识划分成不同的层次和类型,采用框架网络数据结构和产生式规则表示地理学领域专家的专门性知识与经验性知识,使知识的表示简洁易于理解。在知识的推理策略中引入不精确推理机制,保证了推理机制的实现。以乌鲁木齐河流域为试验区,阐述了流域土地合理利用规划决策专家知识的表示方法,以及专家知识表示的基本规则和系统专家知识的组织方式。此外,还讨论了系统推理机推理规则的设计原理与组织方法及推理算法。     

地理专家知识表示的框架网络模型研究

基于框架网络结构模型的专家知识表示方法 ,采用知识的框架网络结构描述地学环境的实体单元 ,将各级专家知识的表示以指针连接 ,形成了由知识到语义的专家知识表示的框架网络。以乌鲁木齐河流域土地合理利用规划决策专家系统的构建为例 ,阐述了该模型的构造方法和实现方案 ,并讨论了系统知识库的组织结构和推理规则的设计原理。     

An intelligent expert system shell for knowledge-based Geographical Information Systems: 1. The tools

Abstract

An intelligent expert system shell for the development of knowledge-based geographical information systems (GIS) is examined in this two-part article. Basic concepts and the overall architecture of the shell are discussed in the present part. Fuzzy logic and expert systems technology are demonstrated to be appropriate methods for approximating human reasoning and enhancing the level of intelligence in GIS. The shell can be employed as an effective and efficient tool for developing knowledge-based GIS.     

地理专家系统的应用现状及发展趋势综述

地理专家系统是专家系统与地理信息系统结合的产物。该文介绍了地理专家系统的概念、框架结构等基础内容，从地理知识库、空间知识发现、推理机制、G1S与ES集成四方面阐述了地理专家系统的研究现状及进展，提出地理专家系统存在的问题及发展趋势。     

Developing a geographic expert system for regional mapping of volcanogenic massive sulfide (VMS) deposit potential

A personal computer-based geographic information system (GIS) is used to develop a geographic expert system (GES) for mapping and evaluating volcanogenic massive sulfide (VMS) deposit potential. The GES consists of an inference network to represent expert knowledge, and a GIS to handle the spatial analysis and mapping. Evidence from input maps is propagated through the inference network, combining information by means of fuzzy logic and Bayesian updating to yield new maps showing evaluation of hypotheses. Maps of evidence and hypotheses are defined on a probability scale between 0 and 1. Evaluation of the final hypothesis results in a mineral potential map, and the various intermediate hypotheses can also be shown in map form.The inference net, with associated parameters for weighting evidence, is based on a VMS deposit model for the Chisel Lake deposit, a producing mine in the Early Protoerzoic Snow Lake greenstone belt of northwest Manitoba. The model is applied to a small area mapped at a scale of 1:15,840. The geological map, showing lithological and alteration units, provides the basic input to the model. Spatial proximity to contacts of various kinds are particularly important. Three types of evidence are considered: stratigraphic, heat source, and alteration. The final product is a map showing the relative favorability for VMS deposits. The model is implemented as aFortran program, interfaced with the GIS. The sensitivity of the model to changes in the parameters is evaluated by comparing predicted areas of elevated potential with the spatial distribution of known VMS occurrences.     

Qualitative spatial reasoning: cardinal directions as an example

ABSTRACT

Abstract. Geographers use spatial reasoning extensively in large-scale spaces, i.e., spaces that cannot be seen or understood from a single point of view. Spatial reasoning differentiates several spatial relations, e.g. topological or metric relations, and is typically formalized using a Cartesian coordinate system and vector algebra. This quantitative processing of information is clearly different from the ways human draw conclusions about spatial relations. Formalized qualitative reasoning processes are shown to be a necessary part of Spatial Expert Systems and Geographical Information Systems.

Addressing a subset of the total problem, namely reasoning with cardinal directions, a completely qualitative method, without recourse to analytical procedures, is introduced and a method for its formal comparison with quantitative formula is defined. The focus is on the analysis of cardinal directions and their properties. An algebraic method is used to formalize the meaning of directions. The standard directional symbols (N, W, etc.) are supplemented with a symbol corresponding to an undetermined direction between points too close to each other which greatly increases the power of the inference rules. Two specific systems to determine and reason with cardinal directions are discussed in some detail.

From this example and some other previous work, a comprehensive set of research steps is laid out, following a mathematically based taxonomy. It includes the extension of distance and direction reasoning to extended objects and the definitions of other metric relations that characterize situations when objects are not disjointed. The conclusions compare such an approach with other concepts.     

地理专家系统的试验研究

80年代人工智能与地理学的结合产生了地理专家系统(GES)。地理专家系统是继地理信息系统之后,对地理知识与推理形式化,建立的具有一定地理专业水平的计算机软件系统。地理专家系统由地理数据库,地理知识库和地理推理机三大部分组成。地理数据库是GES的数据管理模块,我们设计了一套地理信息系统MCGIS来作为微机地理专家系统MCGES的综合数据库;地理知识库是GES的知识管理模块;地理推理机则是GES的推理控制策略的程序体现。本文结合一个用C语言自行设计的基于规则的产生式微机地理专家系统MCGES阐述了地理专家系统的原理和方法。     

A qualitative model for the simulation of traffic behaviours in a multi-lane environment

Qualitative modelling of spatial relationships has often been considered as a contextindependent task that aims at a reasoning model in generic form. Despite the primary interest in these models, there is still a sufficiently large scope for context-dependent reasoning in space and time. This paper proposes a qualitative spatial reasoning model, oriented to the modelling and simulation of several cars acting in a multi-lane circuit, which can be considered as an illustrative example of a constrained frame of reference. The modelling objects of interest are individual cars whose cardinal relationships to external cars and actions are modelled. This dynamic system is analysed, and a set of interrelationships is identified at different levels of abstraction, together with inference rules that model the displacement of several cars in a circuit. The potential of this model is illustrated and calibrated using an agent-based prototype.     

FUZZY MULTIVARIATE RULE-BUILDING EXPERT SYSTEMS:MINIMAL NEURAL NETWORKS

A fuzzy multivariate rule-building expert system (FuRES) has been devised which also functions as aminimal neural network.This system builds rules from training sets of data that use featuretransformation in their antecedents.The rules are constructed using the ID3 algorithm with a fuzzyexpression of classification entropy.The rules are optimal with respect to fuzziness and can accommodateoverlapped and underlapped clusters of data.The FuRES algorithm combines the benefits obtained fromsimulated annealing and gradient optimization,which provide robustness and efficiency respectively.FARES classification trees support OR logic in their inference.The system automatically generatesmeaningful and consistent certainty factors during rule construction.Unlike other neural networks,FuRES uses local processing which furnishes qualitative information in the rule structure of itsclassification trees and variable loadings of the weight vectors.     

土壤类型遥感识别推理决策器研究

介绍了干旱区土壤类型遥感识别推理决策器的设计原理与实现方法。在用TM遥感图像对土壤类型进行非监督分类的基础上，建立了正向推理与逆向推理相结合的推理机制，对土壤类型进行分类识别决策。用知识表示的产生式规则与框架式规则相结合的数据结构表示土壤学专家的土壤分类识别知识。用像结构模式建立了土壤分类识别的规则，构造了土壤分类判决树，并用典型像例模式进行了各类型土壤判据文件的组织。用该方法对新疆天山北麓阜康试验区的土壤分类识别进行了试验研究。结果表明，该方法分类精度可靠，为干旱区土壤分类识别开辟了一条新的途径。     

地理案例推理及其应

由于案例推理接近于人类认识、解决问题最原始的思维方式，具有在无法获取机理模型、确定规则或统计模型时，采用简单的历史相似性实现问题的定量求解和预测的特点，成为当前人工智能中备受关注的领域。在讨论案例推理方法发展的基础上，探讨应用案例推理进行地学问题求解的具体方法。从地学数据分析的角度提出了地理案例推理，并结合地理案例的特点，具体讨论了地理案例的表达模式和推理模型，最后结合东海中心渔场预报的实际工作，给出这种方法的具体应用实例。     

数字地形分析应用适配性知识的案例表达与推理方法

数字地形分析(Digital Terrain Analysis, DTA)在应用时依赖于建模知识,尤其是关于所建的应用模型是否与研究区特点、数据等条件相适配的知识(称为“应用适配性知识”);由于这类知识难以形式化表达,现有的数字地形分析工具对此类知识缺乏利用,从而导致普通用户在应用数字地形分析时建模困难。针对该问题,设计了一套数字地形分析领域应用适配性知识的案例表达与相应的推理方法。以美国32个河网提取案例为例,通过交叉验证,初步表明案例及其推理应用方法适合于数字地形分析领域应用适配性知识的形式化表达与应用,该方法通过与建模环境的集成,可大幅降低数字地形分析应用建模难度。     

Spatial inference with an interchangeable rule format

Rule interchange between information systems is expanding as new interoperable rule formats are emerging from research. However, existing spatial inference systems generally operate on locally stored data with an internal rule format. Consequently, their design offers little support or facilities for rule interchange. This article presents the requirements, components and design for a spatial inference system with rule interchange. Computational efficiency and overall functionality of the design are considered separately, with the latter demonstrated using encoded agricultural legislation and data. A spatial inference system with rule interchange is based on three primary components: rule representation, spatial functionality and data integration. Of these, the interoperable rule representation and data integration distinctly differ from existing spatial inference systems. The presented inference system combines a spatial superset of the W3C Rule Interchange Format (RIF) with full Open Geospatial Consortium simple feature access (OGC SFA) functionality and on-demand data integration utilising Resource Deception Framework (RDF). The design was found to be effective with a computational efficiency depending predominantly on the spatial operations. This design could be further adapted to implement spatial extensions for existing inference systems. Considerable benefits were also discovered when RIF was used as the native language for the inference engine, thereby removing the need for rule transformations and facilitating on-demand data integration with the GML.     

灾害地貌专家系统的结构设计

灾害地貌专家系统采用压缩编码方式存贮各种地学专题图形和遥感图象数据,具有数据与图象的存贮更新、查询检索、分析处理、图象显示和自动制图功能。该系统可以对灾害地貌过程进行专家级的预测和评价,并对灾害地貌的综合治理与区域规划提出几种可行性方案供用户选择。用该系统对天山职拉沟流域泥石流灾害地貌进行了试验研究,并对试验结果进行了系统评价。     

Framework for probabilistic geospatial ontologies

Partial knowledge about geospatial categories is important for practical use of ontologies in the geospatial domain. Degree of overlaps between geospatial categories, especially those based on geospatial actions concepts and geospatial enitity concepts, need to be specified in ontologies. Conventional geospatial ontologies do not enable specification of such information, and this presents difficulties in ontology reasoning for practical purposes. We present a framework to encode probabilistic information in geospatial ontologies based on the BayesOWL approach. The approach enables rich inferences such as most similar concepts within and across ontologies. This paper presents two case studies of using road‐network ontologies to demonstrate the framework for probabilistic geospatial ontologies. Besides inferences within the probabilistic ontologies, we discuss inferences about most similar concepts across ontologies based on the assumption that geospatial action concepts are invariable. The results of such machine‐based mappings of most similar concepts are verified with mappings of concepts extracted from human subjects testing. The practical uses of probabilistic geospatial ontologies for concept matching and measuring naming heterogeneities between two ontologies are discussed. Based on our experiments, we propose such a framework for probabilistic geospatial ontologies as an advancement of the proposal to develop semantic reference systems.     

An event-based methodology for climate change and human–environment research

Abductive causal eventism (ACE) is an analytical methodology based on a pragmatic view of research methods and explanation that places at the center of research inquiry the answering of ‘why’ questions about events, including human actions or environmental changes of interest. When used in human–environment research, the methodology entails constructing causal histories of interrelated social and/or biophysical events backward in time through a process of eliminative inference and reasoning from effects to causes, called abduction. ACE encourages an eclectic use of methods, models, and theoretical ideas. It fosters integrative, interdisciplinary analysis without being committed either to systems as ontological entities or to holistic analytical frameworks. By not privileging particular theories or explanatory factors in advance, ACE enables researchers to interrogate the plausibility of different causal influences, including local environmental changes that may be related to much wider changes in climate. Likewise, behaviors or practices that may be of value in light of anticipated environmental changes can be studied without their presuming to be caused by these changes or by changes in climate per se. Research on coastal mangrove planting for storm protection in the Philippines and upland tree planting in St. Lucia are used to illustrate these arguments.     

Computer Expert System for Epithermal Gold-Silver Deposit Prognostication (Okhotsk–Chuckchee Volcanic Belt,Northeast Russia)

The Okhotsk–Chuckchee volcanic belt (OCVB) represents a world class gold–silver province, being a well-studied part of the Circum-Pacific Rim. Extensive research, preceding the system compilation, included collection of various data on major gold–silver, silver–lead–zinc, and tin–silver ore fields (with some data on minor gold–quartz, gold–sulfide, and gold–rare-metal sites). The database consists of 129 ore fields of the OCVB and comprises 433 geological, mineralogical, geochemical, and geophysical indicators. Using the original pattern recognition programs (accounting the frequencies and correlation of the indicators), the most informative indicators were selected, their weights (both positive and negative) were identified, and an expert system (solution tree) was compiled. The computer system is intended to be an assistance to an expert in the ore field potential evaluation. The system includes several blocks of questions with a goal of classifying the geochemical type and size of the analyzed ore field. The questions have an indirect character, so they could be answered by an expert at early stages of exploration (before drilling). The level of recognition of sizes of the ore sites in the control populations is 95% or higher. The aim of the expert system is to reestimate a large number of known ore sites in a relatively brief period of time with the aid of the personal computer. The revealed patterns could be used in other less studied parts of the Circum-Pacific belt.     

基于空间推理的巷道三维拓扑建立

基于空间关系和空间推理理论,结合MGIS领域的需求建立巷道三维拓扑网络模型,研究巷道三维空间关系推理的算法及业务逻辑实现;开发巷道三维拓扑自动生成系统,实现了巷道空间与属性数据管理、三维拓扑自动处理、多尺度成果输出及基于三维拓扑的路径查询,为基于巷道的通风、运输、避灾、多尺度变换等应用奠定了基础。     

Nearness relations in environmental space

The notion of proximity is a fundamental component of any comprehensive ontology of space. This paper presents an experiment with human subjects concerning the vague spatial relation 'near' in environmental space. After the topic is introduced and relevant previous work surveyed, the experiment is described. Three approaches to experimental analysis are presented and discussed: nearness neighbourhoods as regions with broad boundaries, fuzzy nearness and distance measures, and four-valued logic. Issues discussed in further detail are the truth gap-truth glut hypotheses regarding the psychology of vague predicates, and formal properties of the three-valued nearness relation. The conclusion is drawn that formal theories can usefully underpin and be applied to reasoning with vague spatial notions, such as nearness.