基于TF-IDF和GIoVe算法面向多种类别文本自动分类系统的优化研究

打开文本图片集

关键词 TF-IDF算法；GloVe模型；文本自动分类；关键词位置；词性；语义扩展

分类号 G251;G258.6

DOI10.16810/j.cnki.1672-514X.2025.010.006

AbstractEficientlyorganizingandautomaticallcategorizing textthrough keywordretrievaland specifyingone or more categorylabelsisanimportantwaytodiscoverimplicitrelationshipsindocuments，promoteknowledgedissemination，and driveinnovation.However，the prominentissuesafectingthe eficientand preciseautomaticclasificationof documents are the lack of semantic information inkeywordsandthepoor accuracyof keyword recognition due to factors such as the acquisitionlocation，partfspeech，andcomprehensivenessofslectedkeywords.Inlightofthis，this paperconstructstext automatic clasification system thatcombines TF-IDF （Term Frequency-Inverse DocumentFrequency）and GloVe （Global Vectorsfor WordRepresentation）.First，thesystemimproves theTF-IDFalgorithmbyconsideringpart-of-speechnfluence factorsand positionweightcoeffcients，inordertocompensate forthedeficienciesofthetraditionalTF-IDFalgorithm in keyword identificationandsemanticanalysis.Second，itfurther expands the keywordsetusing the GloVe model，chieving an accuracyandrecallateof92.6and9O.9 fortextclassification.Finally，through experimentalcomparisons，theuperrity and effectiveness of this system in handling multi-category text classification tasksare validated.

KeyWordsTF-IDFalgorithm.GloVemodel.Automatictextlasification.KeyordpositionPartof-speec.Semanticexpansion.

0 引言

通过检索关键词指定一个或多个类别标签，从而实现文本的高效组织和自动分类，是发现文档中的隐含关系并推动知识传播和创新的重要途径。（剩余9590字）