Product Attribute Extraction Based on Affinity Propagation Clustering Algorithm and Pointwise Mutual Information Pruning

With the rapid development of e-commerce platforms, how to quickly sort out and summarize a large number of product reviews on e-commerce websites has become an urgent problem. Product attribute extraction is one of the methods to solve this problem. Automatic extraction of product attributes is one...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:2019 International Conference on Artificial Intelligence and Advanced Manufacturing (AIAM) s. 662 - 666
Hlavní autoři: Fang, Yiqiu, Li, Chunjiang, Ge, Junwei
Médium: Konferenční příspěvek
Jazyk:angličtina
Vydáno: IEEE 01.10.2019
Témata:
Affinity propagation
affinity propagation clustering algorithm
attribute extraction
Data mining
Electronic commerce
Filtering
Mutual information
Natural language processing
Noise
pointwise mutual information
Reviews
Syntactics
Vectors
word vector
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:With the rapid development of e-commerce platforms, how to quickly sort out and summarize a large number of product reviews on e-commerce websites has become an urgent problem. Product attribute extraction is one of the methods to solve this problem. Automatic extraction of product attributes is one of the important research contents in the field of natural language processing. A product attribute extraction method based on affinity propagation clustering algorithm and pointwise mutual information pruning is proposed to extract product attributes comprehensively and rapidly. Firstly, candidate product attribute sets are obtained from the corpus according to predefined syntactic rules. Then, the candidate product attribute sets are represented by word vectors and calculating the similarity of attribute words to cluster the candidate attribute sets. Thus the candidate attribute clusters are obtained. At last, the effectiveness of the method is proved in a Chinese product review.
DOI:10.1109/AIAM48774.2019.00137