›› 2016, Vol. 28 ›› Issue (8): 37-42.

• 应急管理专辑 • 上一篇    下一篇

基于多维情景空间表达的两层案例检索算法研究

钱静1, 刘艺2, 刘呈1, 焦玉莹1   

  1. 1. 清华大学公共安全研究院工程物理系, 北京 100084;
    2. 中国人民公安大学治安学院, 北京 100038
  • 收稿日期:2016-05-18 出版日期:2016-08-28 发布日期:2016-09-24
  • 作者简介:钱静,清华大学工程物理系公共安全研究院科研助理,硕士研究生;刘艺,中国人民公安大学治安学院讲师;刘呈,清华大学工程物理系公共安全研究院博士研究生;焦玉莹,清华大学工程物理系公共安全研究院硕士研究生
  • 基金资助:

    国家自然科学基金项目(91324022;91224008;91024032;70601015;70833003);国家科技支撑计划(2015BAK12B01)。

A Two-stage Case Retrieval Method Based on Multi-dimensional Scenario Space Method

Qian Jing1, Liu Yi2, Liu Cheng1, Jiao Yuying1   

  1. 1. Institute of Public Safety Research, Department of Engineering Physics, Tsinghua University, Beijing 100084;
    2. Pubic Order School, People's Public Security University of China, Beijing 100038
  • Received:2016-05-18 Online:2016-08-28 Published:2016-09-24

摘要:

针对突发事件的响应决策,借鉴以往发生的同类型案例应急处置决策经验是一种比较良好的决策处置方法,另一方面,突发事件案例的复杂性造成案例利用困难。将案例推理方法运用到对复杂突发事件的研究非常重要。本文在设计两层案例检索结构的基础上设计案例检索算法,以提高检索及匹配精度,实现高效的案例检索以用于突发事件的决策支持。

关键词: 案例推理方法, 多维情景空间方法, 案例表达, 案例检索

Abstract:

In response to emergencies, learning from those previous similar cases and experiences is a relatively good decision-making method for emergency management. On the other hand, in light of the complexity of disaster cases which may contain lots of scenarios and attributes in different stages of disaster development, it's very difficult to utilize these cases for real situations without subtle processing. The application of case-based reasoning method to the study of complex emergencies is of great importance. Recent researches have proved that scenario-based reasoning is much more efficient than case-based reasoning. This paper develops case retrieval algorithms on optimized hierarchical case retrieval structure. In order to improve the accuracy and efficiency of case retrieval and decision-making process, approaches proposed in this paper are all centering on the concept of scenarios. Parts of a real case are selected to verify the effect of these algorithms.

Key words: case-based reasoning method, multi-dimensional scenario space method, case representation, case retrieval