地学境本协同语义驱动的矿山企业生产知识服务模式构建研究

Considering the imminent requirement of integrating multi-sourced heterogeneous information, mining tacit knowledge and storing knowledge in the environment of big data, knowledge service pattern with cooperation from three areas-data , knowledge and application , are established for production department of mine enterprise under different semantic context. The structure of hybrid ontology that balances global sharing and local context and the modeling method of "building global ontology by experts then auto-extracting local ontology from data resource"are presented through analysis of semantic conflicting for mine production. The shareable approach are raised for describing data of spaces over and under the land and the algorithm is designed for auto-extracting local ontology from multi-sourced data, which implements mapping and communicating between data and semantic knowledge of production department of mine enterprise. And on this basis, semi-automation method of mine production hybrid ontology with context is adopted to establish semantic environment of heterogeneous data for mine enterprise production department according to the experiences of expert. Furthermore, based on actual data, under this driven by semantic environment, the algorithms of auto-implementing data warehouse, knowledge discovery using association analysis are discussed and the stored pattern of massive semantic files are given in order to extract association rule and normalize knowledge storage of mine production. Besides, it is convenient for multi-factor decision-making for department head and knowledge acquisition for workers of mine production. Therefore, this project research is focused on advanced mode of knowledge service system and gives full play to the positive role of knowledge in strategic management.

针对大数据环境下矿山企业多源异构信息整合、隐性知识挖掘和知识存储迫切需求，以矿山生产业务为主题，构建差异情境下数据区、知识区和应用区协同知识服务模式。通过矿山生产地学语义冲突分析，提出兼顾语义全局共通和局部特色双平衡的矿山生产混合本体结构及先全局专家构建后局部数据源自动语义抽取策略。并研究地上下空间数据共享描述方式以及语义抽取算法，实现多源数据与语义知识映射互通。在此基础上，结合专家咨询，交互半自动构建负载情境的矿山生产领域知识本体，完成矿山企业生产部门异构数据语义环境搭建。进而以现有矿山实际生产数据为依托，探讨此语义环境驱动下面向矿山生产决策的数据仓库自动构建、关联分析知识发现算法以及大规模语义多维存储模式，完成矿山企业生产量关联知识规则抽取与规范化存储，便于生产部门领导多因素综合决策及生产基层员工知识获取，搭建矿山企业先进知识服务体系，充分发挥知识资源在企业战略管理中支撑作用。

针对大数据环境下矿山企业多源异构信息整合、隐性知识挖掘和知识存储迫切需求，以矿山生产业务为主题，提出了一种语义全局共享与局部特色双兼顾的地学领域本体半自动构建策略。顾及复杂地学领域多尺度多时相等区域情境特色，在地学本体元组模型的基础上，负载现实地学数据元情境，并将地学事件并入语义模型，搭建境本协同语义静动共存完备模型，增强地学语义构建的准确性与形式化描述的完备性。在此基础上，依托多维概念格、模糊概念格以及加权概念格理论，通过对数据集的预处理与多维序列构建，生成能够清晰描述多维概念、属性、关系的多维概念格层次结构，实现面向多形式、多载体和多维度的结构化地学空间数据概念对象的快速抽取，提升应对多维地学本体建构效率，同时构建变尺度概念格分块聚合算法和情境敏感概念格胶合本体融合算法。进而结合专家咨询，交互半自动完成矿山企业生产部门异构数据语义环境搭建。结合实际矿山信息系统应用数据，构建融入知识概念的矿山生产数据仓库多维本体模型，探讨依托关系数据库、RDF和图数据库多渠道知识存储模式。对原有矿山生产系统进行重构和升级，拓展开发矿山企业生产多维关联知识分析以及共享知识实体识别应用模块，助力于矿山生产部门领导多因素综合决策，提供矿山生产基层员工专业知识学习平台，实现矿山企业隐性知识的显性化，形成矿山生产业务应用与知识学习一体化服务模式，充分发挥了知识资源在企业战略管理中支撑作用。