越览（130）——精读期刊论文的2方法原理（2）

摘要：This issue will introduce the principle of the method of the intensively read replica paper "Crowd intelligence knowledge mining m

分享兴趣，传播快乐，

增长见闻，留下美好。

亲爱的您，这里是LearingYard学苑！

今天小编为您带来“越览（130）——精读期刊论文

《基于共词网络的群智知识挖掘方法

——在应急决策中应用》的

2方法原理（2）”。

欢迎您的访问！

Share interest, spread happiness,

increase Knowledge, and leave beautiful.

Dear, this is the LearingYard Academy!

Today, the editor brings the

＂Yue Lan (130)：Intensive reading of the journal article

＇Crowd intelligence knowledge mining method

based on co-word network– application

in emergency decision-making’

2 Principle of the method(2)".

Welcome to visit!

一、内容摘要（Summary of Content）

本期推文将从思维导图、精读内容、知识补充三个方面介绍精读复刻论文《基于共词网络的群智知识挖掘方法——在应急决策中应用》的2方法原理（2）。

This issue will introduce the principle of the method of the intensively read replica paper "Crowd intelligence knowledge mining method based on co-word network – application in emergency decision-making" in terms of mind maps, intensively read content, and knowledge supplementation.

二、思维导图（Mind mapping）

三、精读内容（Intensive reading content）

（一）群智知识价值测度（Crowd-wisdom knowledge value measurement）

本节介绍了大数据时代，社交网络UGCs是获取公众情感与诉求的重要来源，但质量参差不齐，影响决策可靠性。为降低不确定性，需对群智知识价值进行测度。研究表明，UGCs主题信息越丰富，用户对危机事件的认知越全面，数据价值越高。本文借鉴COVID-19疫情期间微博数据研究，基于TF-IDF量化关键词重要性，以评估信息价值。

This section introduces that in the era of big data, social network UGCs are an important source of public sentiment and appeals, but the quality is uneven, which affects the reliability of decision-making. In order to reduce uncertainty, it is necessary to measure the value of crowd intelligence knowledge. Studies have shown that the richer the topic information of UGCs, the more comprehensive the user's understanding of the crisis event, and the higher the data value. This article draws on the research on Weibo data during the COVID-19 epidemic and quantifies the importance of keywords based on TF-IDF to evaluate the value of information.

本文提出基于关键词和博文主题价值的群智知识测度方法。首先，通过共词网络分析计算关键词的加权度中心度，以衡量其在社区中的主题价值。然后，结合博文内关键词的TF-IDF值及其主题价值，测度博文的整体主题价值。进一步地，综合博文的主题价值和社交网络影响力，构建单调递增函数以衡量博文的综合价值。最后，结合关键词主题价值和博文价值，计算主题社区的整体强度，实现对群智知识价值的测度。

This paper proposes a method for measuring crowd intelligence knowledge based on keywords and blog post topic values. First, the weighted degree centrality of keywords is calculated through co-word network analysis to measure their topic value in the community. Then, the overall topic value of the blog post is measured by combining the TF-IDF value of the keywords in the blog post and their topic value. Furthermore, a monotonically increasing function is constructed to measure the comprehensive value of the blog post by combining the topic value of the blog post and the influence of the social network. Finally, the overall strength of the topic community is calculated by combining the keyword topic value and the blog post value to achieve the measurement of crowd intelligence knowledge value.

（二）基于群智知识生成动态决策属性信息（Generate dynamic decision attribute information based on crowd intelligence knowledge）

本节介绍了特大公共安全突发事件发生时，需要迅速召集专家组进行多阶段、多属性的群体决策，以制定最佳应对策略。与此同时，公众在社交媒体上的行为大数据也在同步生成，既反映了应急决策的反馈，也体现了对后续决策的诉求，为危机处置提供支持。然而，仅依靠公众数据难以确保决策的科学性和时效性，因此专家组仍是应急决策的核心。由于专家在知识、经验、风险偏好等方面存在差异，他们对同一事件的应对策略可能不同。为更好地结合专家智慧与群智知识，本文提出基于群智知识的动态决策属性信息生成方法，利用公众行为数据为专家提供实时参考，优化应急决策的科学性和有效性。

This section introduces the need to quickly convene an expert group to conduct multi-stage, multi-attribute group decision-making to develop the best response strategy when a major public safety emergency occurs. At the same time, the public's behavioral big data on social media is also generated synchronously, which not only reflects the feedback of emergency decision-making, but also reflects the appeal for subsequent decision-making, providing support for crisis management. However, it is difficult to ensure the scientificity and timeliness of decision-making by relying solely on public data, so the expert group is still the core of emergency decision-making. Due to differences in knowledge, experience, risk preferences, etc., experts may have different response strategies for the same incident. In order to better combine expert wisdom with crowd wisdom knowledge, this paper proposes a dynamic decision attribute information generation method based on crowd wisdom knowledge, using public behavior data to provide real-time reference for experts, and optimize the scientificity and effectiveness of emergency decision-making.

在应急决策中，专家群体负责确定决策属性权重。首先，基于群智知识提取的主题强度进行规范化，计算客观属性权重，并按权重降序排列。其次，专家群体通过成对比较法评估决策属性重要性，形成个体偏好矩阵，并利用加权平均方法聚合，得到群体偏好矩阵，从而计算主观属性权重并排序。最后，为提高权重合理性，引入一致性测度，结合客观与主观权重，通过优先级比较确定最终的决策属性权重，使其更加科学合理。

In emergency decision-making, the expert group is responsible for determining the weights of decision attributes. First, the topic strength extracted based on the crowd intelligence knowledge is normalized, the objective attribute weights are calculated, and they are arranged in descending order of weight. Secondly, the expert group evaluates the importance of decision attributes through the pairwise comparison method to form an individual preference matrix, and aggregates it using the weighted average method to obtain the group preference matrix, thereby calculating the subjective attribute weights and sorting them. Finally, in order to improve the rationality of the weights, the consistency measure is introduced, combining the objective and subjective weights, and determining the final decision attribute weights through priority comparison, making it more scientific and reasonable.

四、知识补充（Knowledge supplement）

Rank Aggregation（排序聚合）方法是一种用于整合多个排序列表的方法，旨在在多个专家或数据来源提供的排序结果之间找到最合理的综合排序。它广泛应用于群体决策、信息检索、机器学习和推荐系统等领域。

Rank Aggregation is a method for integrating multiple ranking lists, aiming to find the most reasonable comprehensive ranking among the ranking results provided by multiple experts or data sources. It is widely used in group decision-making, information retrieval, machine learning, and recommendation systems.

（一）基本思想（Basic idea）

在多个排序结果存在时，Rank Aggregation 方法尝试找到一个全局排序，使其尽可能接近所有输入排序的整体共识。例如，在应急决策专家群体评估多个决策属性的重要性时，每位专家可能给出不同的排序结果，Rank Aggregation 方法能帮助整合这些个体排序，以生成一个合理的全局排序。

When multiple ranking results exist, the Rank Aggregation method attempts to find a global ranking that is as close as possible to the overall consensus of all input rankings. For example, when a group of emergency decision-making experts evaluate the importance of multiple decision attributes, each expert may give a different ranking result. The Rank Aggregation method can help integrate these individual rankings to generate a reasonable global ranking.

（二）常见的方法（Common methods）

1. 波达计数法（Borda Count）

每个候选项（属性）根据其在各个排序列表中的排名得分，得分较高者最终排名靠前。

Each candidate (attribute) is ranked according to its ranking score in each sorted list, and the one with the higher score is ranked higher in the end.

2. 凯门尼-杨方法（Kemeny-Young Method）

寻找一个全局排序，使其与所有个体排序的偏差（逆序对数量）最小。

Find a global sorting that minimizes the deviation (number of inversion pairs) from all individual sortings.

计算方式：计算所有可能排序的逆序对数量，选择逆序最少的排序作为最终排序。

Calculation method: Calculate the number of inversion pairs of all possible sortings, and select the sorting with the least inversions as the final sorting.

3. 中位数排序聚合（Median Rank Aggregation）

取所有专家给出的排名的中位数，作为最终排序。

Take the median of the rankings given by all experts as the final ranking.

适用于少数极端意见可能影响整体排序的情况。

Applicable to situations where a few extreme opinions may affect the overall ranking.

4. 基于马尔可夫链的方法（Markov Chain Based Methods）

将排序问题建模为马尔可夫过程，基于候选项的“胜率”进行迭代计算，最终得到一个稳定排序。

The sorting problem is modeled as a Markov process, and iterative calculations are performed based on the "winning rate" of the candidate items to eventually obtain a stable sorting.

（三）在应急决策中的应用（Application in emergency decision making）

在应急决策中，专家可能根据自身经验和专业知识对多个决策属性进行成对比较，形成不同的排序列表。使用 Rank Aggregation 方法，可以综合多个专家的意见，生成一个合理的全局排序，从而确定各属性的最终权重，提高决策的科学性和一致性。

In emergency decision-making, experts may compare multiple decision attributes in pairs based on their own experience and expertise, forming different ranking lists. Using the Rank Aggregation method, we can combine the opinions of multiple experts to generate a reasonable global ranking, thereby determining the final weight of each attribute and improving the scientificity and consistency of decision-making.

今天的分享就到这里了。

如果您对文章有独特的想法，

欢迎给我们留言，让我们相约明天。

祝您今天过得开心快乐！

That's all for today's sharing.

If you have a unique idea about the article,

please leave us a message,

and let us meet tomorrow.

I wish you a nice day!

文案|yyz

排版|yyz

审核|hzy