基于LCS的应急决策文本相似性比对分析模型

作为应急管理中的主要工作内容，应急决策在整个应急管理工作中具有战略意义，为了对危机情境下的应急决策主体进行聚类分析，定量计算不同决策主体之间相应决策文本的区别。借鉴最长公共子序列模型（LCS）的原理，从决策文本中提取主题词，根据决策文本的特点将其分为“目标”和“行动”2级，结合样本总体确定替换矩阵、空位罚分规则和目标函数，构建高层应急决策文本相似性比对分析模型；运用Needleman Wunsch算法对该模型进行求解；通过对得到的地震情景下的高层应急决策文本的分析，证明模型的有效性和可行性；为了避免序列长度差过大对聚类分析的结果影响，在此基础上提出相对相似性得分的概念，并通过与VSM算法得到的结果进行比较，证明该模型在应急决策文本相似性分析中的优越性。

Comparative analysis model on similarity of emergency decision making texts based on LCS

As the main content of emergency management, the emergency decision making has strategic significance in the whole emergency management work. In order to carry out the cluster analysis on the decision makers in the crisis situation, and quantitatively calculate the difference between the corresponding decision making texts of different decision makers, taking the principles of longest common subsequence model (LCS) for reference, the subject words were extracted from the decision making texts and divided into two levels of "target" and "action" according to the characteristics of decision making texts. Combined with the sample population, the replacement matrix, vacancy penalty rule and objective function were determined, and a comparative analysis model on similarity of emergency decision making texts by the senior people was constructed. The model was solved by using the Needleman Wunsch algorithm, and the effectiveness and feasibility of the model were proved by the analysis on the obtained emergency decision making texts by the senior people under the earthquake scenario. In order to avoid the influence of excessive sequence length difference on the results of cluster analysis, the concept of relative similarity score was proposed on this basis, and the superiority of this model in the similarity analysis of emergency decision making texts was proved by comparing the results with those obtained by VSM algorithm.