Accès libre

Lexical co-occurrence network and semantic relation mining based on English corpus

Guimei Pan

Guimei Pan

Zhanjiang University of Science and TechnologyZhanjiang, China
Recherchez cet auteur sur
Sciendo|Google Scholar
Pan, Guimei
  
17 mars 2025
Lexical co-occurrence network and semantic relation mining based on English corpus's Cover Image
À propos de cet article
Article précédent
Article suivant
Citez
Télécharger la couverture
Publié en ligne: 17 mars 2025
Reçu: 25 oct. 2024
Accepté: 09 févr. 2025
DOI: https://doi.org/10.2478/amns-2025-0209
Mots clés
© 2025 Guimei Pan, published by Sciendo
This work is licensed under the Creative Commons Attribution 4.0 International License.

Figure 1.

Complex network discovery semantic relationship steps
Complex network discovery semantic relationship steps

Figure 2.

General operation process of common analysis
General operation process of common analysis

Figure 3.

Clustering results evaluation
Clustering results evaluation

Figure 4.

Different domain words and the node word distribution statistics
Different domain words and the node word distribution statistics

Some examples of artificial relevance in corpus

Word A Word B Degree of correlation
Imperialism Colonialism 0.472
Shampoo Conditioner 0.399
Overseas Chinese Settlers 0.085
Lion tiger Tiger lion 0.551
Yongding River Lugou Bridge 0.364
First order logic Field of theory 0.261
The middle ages Castle 0.277
Symphony Movement 0.211
Taiwan NT 0.316
Husband’s family Wife’s family 0.387
Hewlett-Packard Printer 0.428
Spring Festival Dumplings 0.395

Multi-path semantic correlation calculation results

Characteristics of the correlation algorithm Grade correlation coefficient
Separate use of the classification diagram 0.42
Reciprocal path 0.23
Depth information weighting 0.39
Information content weighting 0.37
Use the document map separately 0.36
A non-directional diagram of a two-way link 0.25
Link to link separately 0.37
Link to the link 0.31
Forward to the connection and adjust the weight of the parameters 0.41(π = 0.85)
Use the first paragraph instead of English 0.33
Integrated document and classification diagram, and adjust the parameters 0.46(λ = 0.32)
Open test masking(WS353) 0.37

Experimental parameter

Parameter name Parameter value
Maximum iteration number 400
Iteration out of the threshold 0.003
The single document extracts the number of keywords N 10
Damping factor d 0.61
Vertex score 1.0
Slide pane size w 2/6/10/15/20

Keyword extraction comparison

Algorithm Index 2 6 10 15 20 MEAN
TextRank
P 0.3625 0.3212 0.3578 0.3755 0.3964 0.3627
R 0.4689 0.4968 0.4578 0.4772 0.4931 0.4788
F1 0.3715 0.3251 0.3698 0.3745 0.3604 0.3603
I 53 56 59 52 51 54.2
Improved textrank
P 0.4596 0.4685 0.4725 0.4869 0.4911 0.4757
R 0.5521 0.5637 0.5417 0.5698 0.5927 0.5640
F1 0.4122 0.4166 0.4867 0.4474 0.4516 0.4429
I 42 41 38 35 45 40.2

Number of data sets

Classification Text quantity Classification Text quantity
Data set 1 526 Data set 6 64
Data set 2 125 Data set 7 153
Data set 3 348 Data set 8 186
Data set 4 56 Data set 9 54
Data set 5 48 - -
Langue:
Anglais
Périodicité:
1 fois par an
Sujets de la revue:
Sciences de la vie, Sciences de la vie, autres, Mathématiques, Mathématiques appliquées, Mathématiques générales, Physique, Physique, autres
RSS Feed de la revue