基于BERTopic的新兴技术主题识别研究

[目的/意义]识别并预见新兴技术,为企业和政府带来技术先发优势,及时把握技术发展趋势。[方法/过程]本研究采用BERTopic的主题建模方法获取领域主题分布,依据主题向量的余弦相似度合并论文和专利主题,实现新兴主题的识别。[结果/结论]运用BERTopic主题建模方法结合指标评价,能够实现新兴主题与新兴术语的有效识别。以新能源汽车领域为例开展实证研究,使用划分验证期和资料验证法2种方式进行验证,在识别出的16项主题中有12项通过验证,证实了本研究方法的有效性。

基于BERTopic模型的国外信息资源管理研究进展分析

[目的/意义]文章从已有研究成果中提取主题,梳理主要研究方向,展示主题热度变化趋势,为了解和评估国外信息资源管理(IRM)研究发展现状与趋势提供参考。[方法/过程]采用新兴BERTopic模型对2013—2022年期间WoS数据库中IRM相关文献进行主题提取与识别,结合相关主题词及主题距离划分研究方向,并利用动态主题模型揭示国外IRM领域的演变过程。[结果/结论]国外IRM近10年的研究可分为59个主题,可归纳为信息技术及应用、企业信息管理、图书馆管理与服务、健康信息管理、信息用户与服务、IRM基本理论与方法、文献计量与评价7个方向。大多数主题的研究热度变化偏向稳定的趋势,数字化建设、开放数据等部分主题热度逐渐上涨,而外包、知识管理等少数主题热度退却。

Mining Topical Influencers Based on the Multi-Relational Network in Micro-Blogging Sites

In micro-blogging contexts such as Twitter,the number of content producers can easily reach tens of thousands,and many users can participate in discussion of any given topic.While many users can introduce diversity,as not all users are equally influential,it makes it challenging to identify the true influencers,who are generally rated as being interesting and authoritative on a given topic.In this study,the influence of users is measured by performing random walks of the multi-relational data in micro-blogging:retweet,reply,reintroduce,and read.Due to the uncertainty of the reintroduce and read operations,a new method is proposed to determine the transition probabilities of uncertain relational networks.Moreover,we propose a method for performing the combined random walks for the multi-relational influence network,considering both the transition probabilities for intra-and inter-networking.Experiments were conducted on a real Twitter dataset containing about 260 000 users and 2.7million tweets,and the results show that our method is more effective than TwitterRank and other methods used to discover influencers.

基于BERTopic模型的中国能源行业绿色低碳技术主题挖掘与动态演化分析

能源行业绿色低碳转型是实现“双碳”目标的必然要求。依据技术生命周期理论研究中国能源行业绿色低碳技术发展阶段,采用BERTopic模型和余弦相似度计算方法对相关专利进行数据挖掘,分析各个发展阶段的技术主题、技术方向及动态演化规律。结果显示:中国能源行业绿色低碳技术的萌芽期、发展期和成熟期的主题数量分别为6个、12个和15个;萌芽期技术方向较为单一,发展期技术主题涵盖5个技术方向,成熟期涵盖4个技术方向;动态演化过程受到技术进步、政策导向等因素影响而体现出明显的时代特征。

基于BERTopic模型的网络暴力事件衍生舆情探测

[研究目的]在海量用户生成内容中及时探测和剖析网络暴力事件的衍生舆情能够为舆情事件链的演化分析、同类舆情的研判介入、衍生事件的监测预警提供理论支持。[研究方法]使用BERTopic模型对短文本内容主题建模并采用聚类的方式展示主题的潜在层次结构。根据词向量余弦相似度设计主题衍生度的计量算法,同时融合词共现网络在文档-词语层面信息捕捉的优势以及桑基图直观演示舆情演化过程的特点,衡量主题间的影响力与衍生关系。[研究结论]在开源数据集下多组主题模型的对照实验中,BERTopic模型在短文本建模以及下游任务的平均得分提高2.13%。在网络暴力热点事件的应用实例中,多维细粒度分析与交互式可视化方法可达到直观展示暴力事件的主题聚类、词义关联与演化态势的效果,实现网络暴力事件衍生舆情的探测与分析。

基于BERTopic算法的引文主题实证分析——以一篇高被引诺贝尔生理学或医学奖论文为例

[目的/意义]引文主题识别/分析(CTR/CTA)是引文内容分析(CCA)的一项重要研究议题,通过对引文语料中蕴涵主题信息的识别和提取,可望为论文学术贡献评价、知识扩散及演化分析等问题的解决提供新的研究思路。[过程/方法]以一篇高被引诺贝尔生理学或医学获奖关键论文为例,采用BERTopic算法对其引文句语料进行主题识别,并对识别出的引文主题展开多个维度的分析与讨论。[结果/结论]对高被引论文开展引文主题识别分析,有助于更全面细致地揭示其学术贡献内容及演化趋势;BERTopic算法能较好识别案例文献的多个引文主题,且不同引文主题的施引文献特征分布不尽相同;对引文主题重要性、演化趋势及其与原文主题差异性的分析,能多维度刻画研究同行对案例文献学术贡献的认识,表明CTR/CTA研究对学术论文评价具有深入探索价值。

Analyzing topics in social media for improving digital twinning based product development

Digital twinning enables manufacturers to create digital representations of physical entities,thus implementing virtual simulations for product development.Previous efforts of digital twinning neglect the decisive consumer feedback in product development stages,failing to cover the gap between physical and digital spaces.This work mines real-world consumer feedbacks through social media topics,which is significant to product development.We specifically analyze the prevalent time of a product topic,giving an insight into both consumer attention and the widely-discussed time of a product.The primary body of current studies regards the prevalent time prediction as an accompanying task or assumes the existence of a preset distribution.Therefore,these proposed solutions are either biased in focused objectives and underlying patterns or weak in the capability of generalization towards diverse topics.To this end,this work combines deep learning and survival analysis to predict the prevalent time of topics.We propose a specialized deep survival model which consists of two modules.The first module enriches input covariates by incorporating latent features of the time-varying text,and the second module fully captures the temporal pattern of a rumor by a recurrent network structure.Moreover,a specific loss function different from regular survival models is proposed to achieve a more reasonable prediction.Extensive experiments on real-world datasets demonstrate that our model significantly outperforms the state-of-the-art methods.

A Video Captioning Method by Semantic Topic-Guided Generation

In the video captioning methods based on an encoder-decoder,limited visual features are extracted by an encoder,and a natural sentence of the video content is generated using a decoder.However,this kind ofmethod is dependent on a single video input source and few visual labels,and there is a problem with semantic alignment between video contents and generated natural sentences,which are not suitable for accurately comprehending and describing the video contents.To address this issue,this paper proposes a video captioning method by semantic topic-guided generation.First,a 3D convolutional neural network is utilized to extract the spatiotemporal features of videos during the encoding.Then,the semantic topics of video data are extracted using the visual labels retrieved from similar video data.In the decoding,a decoder is constructed by combining a novel Enhance-TopK sampling algorithm with a Generative Pre-trained Transformer-2 deep neural network,which decreases the influence of“deviation”in the semantic mapping process between videos and texts by jointly decoding a baseline and semantic topics of video contents.During this process,the designed Enhance-TopK sampling algorithm can alleviate a long-tail problem by dynamically adjusting the probability distribution of the predicted words.Finally,the experiments are conducted on two publicly used Microsoft Research Video Description andMicrosoft Research-Video to Text datasets.The experimental results demonstrate that the proposed method outperforms several state-of-art approaches.Specifically,the performance indicators Bilingual Evaluation Understudy,Metric for Evaluation of Translation with Explicit Ordering,Recall Oriented Understudy for Gisting Evaluation-longest common subsequence,and Consensus-based Image Description Evaluation of the proposed method are improved by 1.2%,0.1%,0.3%,and 2.4% on the Microsoft Research Video Description dataset,and 0.1%,1.0%,0.1%,and 2.8% on the Microsoft Research-Video to Text dataset,respectively,compared with the existing video captioning methods.As a result,the proposed method can generate video captioning that is more closely aligned with human natural language expression habits.

Vat-based photopolymerization 3D printing:From materials to topical and transdermal applications

Three-dimensional(3D)printing is an innovative manufacturingmethod with the potential to revolutionize topical and transdermal dosage forms.Nowadays,it is established that Vatbased photopolymerization(VP)3D printing technologies offer superior printing efficiency and versatility compared to other 3D printing technologies available on the market.However,there are some limitations that impair their full application in pharmaceutical contexts,such as the lack of a range of biocompatiblematerials for topical and transdermal applications.This reviewarticle explores all types of VP-based 3D printing and discusses the relevance of implementing this kind of technology.We start with a detailed description of the printing process,focusing on the commercial materials available and lab-made resins proposed by different authors.We also review recent studies in this field,which mainly focus on the fabrication of transdermal devices based on microneedle arrays.In the future,it is expected that the manufacturers of 3D printers invest in modifications to the printing apparatus to allow the simultaneous printing of different resins and/or compound types,which will open frontiers to the personalization of treatment approaches.

Pain perception enhancement in consecutive secondeye phacoemulsification cataract surgeries under topical anesthesia

Cataract is the main cause of visual impairment and blindness worldwide while the only effective cure for cataract is still surgery.Consecutive phacoemulsification under topical anesthesia has been the routine procedure for cataract surgery.However,patients often grumbled that they felt more painful during the second-eye surgery compared to the first-eye surgery.The intraoperative pain experience has negative influence on satisfaction and willingness for second-eye cataract surgery of patients with bilateral cataracts.Intraoperative ocular pain is a complicated process induced by the nociceptors activation in the peripheral nervous system.Immunological,neuropsychological,and pharmacological factors work together in the enhancement of intraoperative pain.Accumulating published literatures have focused on the pain enhancement during the secondeye phacoemulsification surgeries.In this review,we searched PubMed database for articles associated with pain perception differences between consecutive cataract surgeries published up to Feb.1,2024.We summarized the recent research progress in mechanisms and interventions for pain perception enhancement in consecutive secondeye phacoemulsification cataract surgeries.This review aimed to provide novel insights into strategies for improving patients’intraoperative experience in second-eye cataract surgeries.

基于BERTopic模型的数字政府治理领域的主题识别与内容分析

[目的/意义]在信息技术创新、政府数字化转型的背景下,中国学术界输出的“数字政府治理”相关研究体量较大,现阶段亟待对其进行科学整理和综述概览,以回应“政府治理现代化”目标对理论研究的新要求。[方法/过程]以2011-2023年间中国知网数据库收录的有关数字政府治理研究的1829篇文献为数据来源,借助深度学习模型BERTopic和文献计量法CiteSpace相互验证分析数字政府治理领域的研究阶段、研究关键词、研究热点主题及内容等,使得结果具有高准确性和强解释性,以科学有效地探测主题取向及特征,展望数字政府治理领域的未来研究方向。[结果/结论]中国数字政府治理研究在过去的10年间发展迅猛,已成为学者探索的热点论域;现有文献研究热点整体上聚焦于价值导向、治理模式进化、技术工具嵌入;重点主题关注的是“政府数据开放共享研究”“基于数字技术的政务服务改革研究”“城乡场域下的数字化治理研究”“数字政府治理水平的评估研究”及“数字化转型下具体治理领域的实践研究”。未来,数字政府治理仍然是一个具有持续性拓展空间的研究领域,将逐步实现数字政府治理的更广场域触达、精准滴灌、多项学科融合和技术实用主义。

基于Citation Topic的领域期刊主题演化研究——以行为医学为例

基于论文层面的Citation Topic分类体系,通过测度行为医学领域SCI及SSCI期刊引文主题变动情况,展现国际行为医学领域研究重点以及相关支撑学科的演化。

Enhancing BERTopic with Pre-Clustered Knowledge: Reducing Feature Sparsity in Short Text Topic Modeling

Modeling topics in short texts presents significant challenges due to feature sparsity, particularly when analyzing content generated by large-scale online users. This sparsity can substantially impair semantic capture accuracy. We propose a novel approach that incorporates pre-clustered knowledge into the BERTopic model while reducing the l2 norm for low-frequency words. Our method effectively mitigates feature sparsity during cluster mapping. Empirical evaluation on the StackOverflow dataset demonstrates that our approach outperforms baseline models, achieving superior Macro-F1 scores. These results validate the effectiveness of our proposed feature sparsity reduction technique for short-text topic modeling.

基于BERTopic的中药治疗眼科疾病的处方用药规律分析

目的分析中医药治疗眼科疾病的处方用药规律,探讨治疗眼科疾病的更多潜在配伍组合。方法收集眼科专家门诊的中医处方并整理筛选,采用Python 3.9.10中BERTopic算法对复发性虹膜睫状体炎的所有处方进行分析,得到药物关键词聚类组合;采用SPSS 25.0对中药处方数据进行层次聚类,比较两种方法的优劣和药物配伍组合的挖掘效果。结果采用BERTopic算法得到核心药物关键词组合累计2种,包括酒地龙、土茯苓、大青叶、葎草等;应用层次聚类得到的药物组合包括葎草、土茯苓、菝葜等。应用BERTopic技术的算法具有不易受噪声数据影响、提高聚类效率、增强对于处方文本语义理解等诸多优势。结论使用基于BERTopic技术的算法在寻找疾病潜在中药配伍中表现良好,潜力较大,能为眼科疾病的中药配伍组合提供更多参考方向。

电信领域通过《专利合作条约》(PCT)在中国申请专利的技术主题识别与演化趋势——基于BERTopic主题模型的分析

为真实反映针对中国电信领域的技术布局与竞争态势,采集电信领域通过《专利合作条约》(PCT)申请进入中国国家阶段的中国、美国、日本、韩国这4个国家2009—2020年PCT专利摘要文本形成数据集,利用BERTopic主题模型分析挖掘提取各数据集的技术主题及其发展演化情况。结果发现,电信领域中国、美国、日本、韩国进入中国国家阶段的PCT专利申请中,中国在小区测量技术,美国在图像处理、波束成形、音频元器件技术,日本在无线电信系统和电子元器件技术,韩国在图像视频编码和无线电信系统等技术方向各占优势,并逐渐交叉重叠。根据研究发现,提出我国电信行业创新主体要继续锚定基础设施建设,拓展应用层面技术开发和加强国际合作的研发建议。

基于BERTopic主题建模的延时现场救护研究趋势与热点分析

目的分析2009~2024年间国际延时现场救护领域的文献,探究主要研究主题及其发展趋势,以期为未来救护策略提供理论支持。方法系统检索PubMed、Embase、Web of Science和中国知网等数据库,筛选并纳入283篇相关文献。运用BERTopic主题建模技术对文献进行主题识别和关键词分析,并进行可视化展示。结果当前研究主要聚焦在“急救策略研究”“智能技术与信息管理”“实战应用”与“政策与理论研究”等4个方面,预测这些领域将持续成为研究热点。结论国际延时现场救护研究正处于快速发展阶段,建议未来研究深入重点领域,开发有效的救护策略,以提升救治效率和伤员生存率。

基于BERTopic模型的组织成员工作投入研究的主题提取

[目的/意义]旨在通过科学计量方法,挖掘“组织成员工作投入”文献资源中蕴藏的主题,为后续研究提供参考和启示。[方法/过程]搜集中国知网学术期刊数据库中2002-2023年相关文献的摘要和发表年份,经过文本预处理,使用BERTopic模型进行主题提取和可视化分析。[结果/结论]国内现有的关于组织成员工作投入的研究可以分为研究内容和研究方法两大主题集群,均表现出多样化态势。主题时序演化分析揭示了组织成员工作投入研究正逐步转向对个体差异和心理健康的关注。未来研究可从新技术的影响、工作投入动态变化以及跨学科合作等方面进行拓展。

基于BERTopic模型的公众心理应激信息表征分析

[目的/意义]在后疫情时代的背景下,探讨心理应激的信息表征对公众的心理健康建设及社会发展具有重要意义。[方法/过程]使用BERTopic模型对问答数据进行匹配和热度分析,探究不同主题下用户的关注内容和关注程度。[结果/结论]用户心理应激信息表征主要集中在生理反应、认知反应、情感反应、行为反应4种类型。本研究从后疫情时代的视角出发探讨心理应激的信息表征,能够以常态化心理应激干预为目标为决策者提出长期心理支持服务建议,推进线上线下并举的心理服务。

基于XLNet和多粒度对比学习的新闻主题文本分类方法

新闻主题文本内容简短却含义丰富,传统方法通常只考虑词粒度或句粒度向量中的一种进行研究,未能充分利用新闻主题文本不同粒度向量之间的关联信息。为深入挖掘文本的词向量和句向量间的依赖关系,提出一种基于XLNet和多粒度特征对比学习的新闻主题分类方法。首先,利用XLNet对新闻主题文本进行特征提取获得文本中词、句粒度的特征表示和潜在空间关系;然后,通过对比学习R-Drop策略生成不同粒度特征的正负样本对,以一定权重对文本的词向量-词向量、词向量-句向量和句向量-句向量进行特征相似度学习,使模型深入挖掘出字符属性和语句属性之间的关联信息,提升模型的表达能力。在THUCNews、Toutiao和SHNews数据集上进行实验,实验结果表明,与基准模型相比,所提方法在准确率和F 1值上都有更好的表现,在三个数据集上的F 1值分别达到了93.88%、90.08%、87.35%,验证了方法的有效性和合理性。

融合LDA-BN的船舶碰撞事故致因分析

为探究船舶碰撞事故致因及其关系,提升航运安全管理水平,研究提出融合狄利克雷分布(Latent Dirichlet allocation,LDA)与贝叶斯网络(Bayesian Network,BN)的船舶碰撞事故致因分析方法。首先,运用LDA主题模型挖掘361份船舶碰撞事故调查报告,提取27个事故致因主题;其次,利用事故树方法厘清调查报告中致因间的影响关系,构建事故致因贝叶斯网络结构,使用期望最大化算法进行贝叶斯网络参数学习,确定各节点的条件概率,构建事故致因贝叶斯网络模型;最后,通过逆向推理分析、最大致因链分析及敏感性分析,找出导致船舶碰撞事故发生的主要致因因素。结果显示:安全管理不到位、疏忽瞭望、事发水域通航环境复杂是引发船舶碰撞事故可能性大的致因,航线保持不当、应急处置不当、违规穿越锚地是导致船舶碰撞事故发生的最敏感致因因素。