面向任务描述的国防科技情报能力评估指标体系构建研究——基于主题挖掘和对比学习方法

Research on the Construction of the National Defence Science and Technology Information Capability Assessment Indicator System for Task Description:Based on Theme Mining and Contrastive Learning Methods

[目的/意义]在大国博弈和科技竞争背景下,情报需求持续增长,国防科技情报的重要性日益凸显。开展国防科技情报能力评估研究,构建相关评估指标体系有助于国防科技情报任务完成。[方法/过程]文章提出一种面向任务描述的国防科技情报能力评估指标体系构建方法,首先,以中文社会科学引文索引数据库中的相关研究为基础语料,运用LDA主题模型获得分类结果以生成评估指标;其次,以国防科技战略情报综合研判类任务为例,基于语义相似度筛选评估指标;最后,计算指标权重,完成评估指标体系构建。[结果/结论]根据不同情报任务的描述构建有所区别的情报能力评估指标体系,研究结果验证了方法的有效性,综合考虑指标来源主题强度和与任务描述相似度的权重更符合国防科技情报任务的需求。

[Purpose/significance]Under the background of great power games and scientific and technological competition,the demand for information continues to grow,and the importance of national defence scientific and technological information is becoming more and more prominent.Conducting research on the assessment of the capabilities of defence scientific and technological information and constructing a relevant system of assessment indicators will help to fulfil the tasks of defence scientific and technological information.[Method/process]The paper proposes a task description-oriented method for the construction of national defence science and technology information capability assessment index system.Firstly,taking the relevant studies in Chinese Social Sciences Citation Index database as the basic corpus,the theme model is used to obtain the classification results of the themes to generate the assessment indexes.Second,taking the tasks of national defence science and technology strategic information synthesis and judgment as an example,the assessment indexes are screened based on the semantic similarity.Finally,the weights of the indexes are calculated,and the construction of the assessment index system is completed.[Result/conclusion]According to the characteristics of different information tasks to build a differentiated information capability assessment index system,the results verify the effectiveness of the methodology,the comprehensive consideration of the strength of the source theme and the similarity of the index weight is more in line with the needs of the completion of the national defence science and technology information tasks.