旅游地理本体与众包旅游信息的时空语义挖掘研究

In the Internet era with Big Data environment, the semantic sharing of data and calculation models and collaboration among scientists, are becoming increasingly significant. On the Internet, tourism information of Crowdsourcing, such as the knowledge entries of tourism, travel notes, landscape photos, and travel routes etc., is the important element of Big Data. Tourism geography is necessary to find scientific methods of mining Crowdsourcing information, and transform the research paradigm to The Fourth Paradigm: Data-Intensive Scientific Discovery. This research introduces the Ontology technology to tourism geography. The core of the research project includes: 1) Based on the existing tourism ontologies and geography ontologies, a tourism geo-ontology model will be created by combining the domain expert knowledge and new concepts of the Internet; 2) The automatic populating methods for tourism geo-ontology and knowledge checking model based on ontology learning will be explored, and the knowledge database will also be updated dynamically; 3) Some models and tools will be created to find the temporal and spatial semantics within tourism information of Crowdsourcing, such as the models for temporal and spatial semantics extracting, scale transformation, and data precision evaluation; 4) Applying to some models of tourism geography, such as travelling flow, human-land relation of tourism, spatial pattern and process of tourism, researchers will test these tourism geo-ontology model and the data mining methods of temporal and spatial semantics. The research project aims at creating the new methods for using Big data, such as the Crowdsourcing data. The research results will provide theoretical basis of integration of data, models and services, and contribute to collaboration among scientists and collaborative computing between different regions. There are also basic significance for researching on tourism spatial pattern, process and mechanism, and for the tourism policy making and evaluation.

大数据环境下，数据与模型的语义共享和科学家群体协作日益重要。互联网中的旅游知识条目、文本、照片、轨迹等"众包旅游信息"是大数据的重要组成部分。旅游地理学有必要寻找科学利用众包信息的方法，适应"数据密集型科学"的计算模式。项目引入本体思想，基于领域专家知识，充分复用已有旅游和地理本体，并吸纳互联网新概念，建立旅游地理本体模型。研究基于本体学习的本体自动填充与知识校验方法，实现本体知识库的持续更新。探索基于旅游地理本体的众包旅游信息时空语义挖掘方法，设计开发时空语义提取、尺度变换和精度评估的模型与工具等。通过典型区域的旅游流动性、人地关系和旅游格局与过程建模对本体模型和挖掘方法进行验证。项目旨在建立旅游地理学应用众包大数据的关键方法，能够为多时空间尺度上的数据、模型与服务集成，科学家群体协作和区域联合计算提供理论基础，对旅游格局、过程与机制研究和旅游政策制定与效益评价等均具有基础意义。

本项目在新地理信息时代背景下，针对微博、文本、照片、轨迹等众包旅游信息的非规则性和时空语义不确定性等问题，从旅游地理本体的概念、内涵和本体构建方法入手，尝试旅游地理本体的填充与校验方法，设计开发众包旅游信息的时空语义提取、尺度变换和精度评估模型，探索基于旅游地理本体的时空语义挖掘与应用。主要研究成果包括：（1）构建了初步的旅游地理本体理论模型，并完成以张家界旅游地为案例的旅游地理本体实例；建立了以旅游文本、位置微博等为主体的系列众包旅游信息语料库，为基于旅游众包信息的旅游地理研究提供基础；（2）设计了旅游文本内容中的热度地名提取、识别与可视化方法；引入LDA主题模型，提出了多特征维度的旅游文本主题分类方法；（3）设计了针对不同数据源、不同空间尺度的旅游关注度计算方法、旅游行为轨迹栅格化方法，完成了从景区到区域旅游地的多尺度实证案例；（4）引入粒计算模型和包含度方法，提出了多尺度旅游文本景观热度粒计算模型，为旅游众包信息的时空尺度变换和数据挖掘奠定了方法学基础；（5）选择古城古镇、山岳型风景区、区域旅游地等不同类型、不同尺度的旅游空间单元，开展了旅游者景观感知、时空行为特征及其演变过程的研究；（6）初步证明了旅游地理本体支撑下的众包旅游信息在旅游者的流动性、人地关系和旅游格局与过程建模方面的应用价值。.旅游众包信息是旅游地理学研究的优质数据源，本研究针对旅游众包信息的旅游地理学应用方法与案例成果，对多时空尺度的旅游地理格局、过程与机制研究具有理论贡献，在旅游地发展状态的实时监测，旅游地发展规划与预测、旅游政策制定与旅游效益评价方面也具有参考意义。