Research on E-Commerce Review Data Mining Based on Differential Evolution Particle Swarm

In order to realize the accurate mining of e-commerce review data, a clustering method based on adaptive clustering center ranking and differential evolution optimization particle swarm algorithm is proposed. In order to improve the compactness and accuracy of clustering results, the adaptive cluste...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:2023 7th Asian Conference on Artificial Intelligence Technology (ACAIT) s. 1040 - 1046
Hlavní autoři: Liang, Jiafu, Li, Meiqiong
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: IEEE 10.11.2023
Témata:
Clustering algorithms
data analysis
Data mining
differential evolution algorithm
E-commerce platform
Electronic commerce
Particle swarm optimization
Reviews
Sociology
user review
Vectors
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:In order to realize the accurate mining of e-commerce review data, a clustering method based on adaptive clustering center ranking and differential evolution optimization particle swarm algorithm is proposed. In order to improve the compactness and accuracy of clustering results, the adaptive clustering center vector ranking method is introduced to sort clustering centers by similarity. Then, differential evolution is utilized to improve the particle swarm algorithm, so as to improve the accuracy of traditional clustering. The results show that when the population size is 30 and K value is 5, the clustering results of the proposed improved differential evolution particle swarm optimization algorithm are the best. The average F value of the proposed algorithm is 0.163, 0.644, 2.449 and 1.532 higher than that of the current methods with better clustering results such as unimproved DEPSO algorithm, GAI-PSO algorithm, IDE-SOM algorithm and ST-CNN algorithm, respectively. The improvement effect is obvious, and the clustering performance is better. The proposed IDEPSO method is adopted to perform clustering analysis on the reviews of e-commerce stores, which shows that buyers pay more attention to the overall design, material quality, comfort, price and logistics. This proves that the clustering of e-commerce reviews can guide the improvement of e-commerce operations, which has certain practical value and is worth promoting.
DOI:10.1109/ACAIT60137.2023.10528392