Production2Vec: a hybrid recommender system combining semantic and product complexity approach to improve industrial resiliency - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2021

Production2Vec: a hybrid recommender system combining semantic and product complexity approach to improve industrial resiliency

Arnault Pachot
Benjamin Albouy-Kissi
Frédéric Chausse

Résumé

COVID-19 health crisis highlights the fragility of European industrial strategies and leads us to develop more agile, distributed, and resilient production models at a territorial level. There are two major challenges in this regard: one is to find solutions to secure supplies and/or industrial value chains, and the other is to identify companies that have the potential to transform their production quickly to cope with an emergency situation. We extended the Word2Vec vector space with products and economic activities allowing us to calculate proximities. We present a methodology based on semantic proximity and productive complexities to assess the ability of an A-company to produce a product B and to anticipate customer/supplier-type collaboration according to industrial quality standards. We consider recommendation topics by intertwining machine learning techniques with semantic approaches, referring to area ontologies incorporating territorial dimensionality. CCS CONCEPTS • Natural language processing; • Recommender systems; • Supply chain management;
Fichier principal
Vignette du fichier
ICAIIS21-5-2.pdf (1013.74 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03276942 , version 1 (02-07-2021)

Identifiants

Citer

Arnault Pachot, Adélaïde Albouy-Kissi, Benjamin Albouy-Kissi, Frédéric Chausse. Production2Vec: a hybrid recommender system combining semantic and product complexity approach to improve industrial resiliency. 2021 2nd International Conference on Artificial Intelligence and Information Systems (ICAIIS ’21), May 2021, Chongqing, China. ⟨10.1145/3469213.3469218⟩. ⟨hal-03276942⟩
53 Consultations
85 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More