China
Africa
Explore chapters and articles related to this topic
Information Extraction
Published in John Atkinson-Abutridy, Text Analytics, 2022
Then a relationship extraction method would attempt to extract semantic relationships from texts that usually occur between two or more named entities. As in the previous examples, relationships can be of different types and are characterized using triplets SVO (Gillon, 2019), that is (Subject,Verb, Object), such as (John, complained, service) for the sentence “John complained about the service”.
An ABGE-aided manufacturing knowledge graph construction approach for heterogeneous IIoT data integration
Published in International Journal of Production Research, 2023
Lei Ren, Yingjie Li, Xiaokang Wang, Jin Cui, Lin Zhang
There are usually a lot of text reports such as zeroing reports, problem handling orders, concession receipt orders, exception release orders during the manufacturing process. How to extract the triples shown as (head entity, relationship, tail entity) to construct the knowledge graph from the vast and complex knowledge base of the manufacturing industry and the attribute information contained in the entity is an essential step in building the knowledge graph the manufacturing industry. The critical problem of knowledge extraction is that how to extract entities, relationships, attributes, and other content from a large amount of semi-structured and unstructured text data. The key technologies are entity extraction and relationship extraction. In the knowledge graph, knowledge is expressed in the form of triple SPO (Subject, Predicate, Object). However, manufacturing knowledge has strong terminology expertise and complex relationships between terms. Traditional entity recognition and relationship extraction methods are challenging to extract manufacturing knowledge effectively. Therefore, extracting knowledge triples from the text report of the manufacturing process is a crucial issue for constructing a large-scale knowledge graph of the manufacturing industry.
A B2B flexible pricing decision support system for managing the request for quotation process under e-commerce business environment
Published in International Journal of Production Research, 2019
K.H. Leung, C.C. Luk, K.L. Choy, H.Y. Lam, Carman K.M. Lee
To identify the hidden relationships in the purchasing behaviour of a particular customer, and the product popularity in the past, processed data from the Data storage and retrieval module is extracted to this module. The data mining algorithm adopted in this system integrates fuzzy sets theory and association rules theory, which had been proposed and used in several previous studies, i.e. Hong, Lin, and Wang (2003), Lau et al. (2009), Ho et al. (2012), and Lee et al. (2015). In general, the use of the fuzzy association rule mining approach can be classified into three stages, namely, (i) parameter setting stage, (ii) relationship extraction stage, (iii) rule evaluation and selection stage. In the application of the fuzzy association rule mining algorithm for extracting hidden relationships, parameters, as well as the fuzzy characteristics of each parameter, are defined and stored in this module. As these parameter settings are crucial to the subsequent data mining process, it is recommended to consult industry experts for defining the fuzzy membership sets and threshold values, such as the confidence values, in accepting the possible association rules. The detailed procedures of applying the fuzzy association rule mining algorithm in generating pricing decision support for suppliers during the RFQ process is depicted in the Case study section.
Entity and relation collaborative extraction approach based on multi-head attention and gated mechanism
Published in Connection Science, 2022
Wei Zhao, Shan Zhao, Shuhui Chen, Tien-Hsiung Weng, WenJie Kang
For two entities, their entity distance refers to the absolute character offset between the last character of the entity that appears first and the last character of the entity that appears second. The distance between related entities influences the effect of relationship extraction, and capturing long-distance entity dependencies is always a difficult problem in relation to extraction. To evaluate the impact of the entity distance on the performance of GANCE, experiments are conducted on the CoNLL04 dataset.