Building Vector Representations for Candidates and Projects in a CV Recommender System (CROSBI ID 696207)
Prilog sa skupa u zborniku | izvorni znanstveni rad | međunarodna recenzija
Podaci o odgovornosti
Kurdija, Adrian Satja ; Afrić, Petar ; Šikić, Lucija ; Plejić, Boris ; Šilić, Marin ; Delač, Goran ; Vladimir, Klemo ; Srbljić, Siniša
engleski
Building Vector Representations for Candidates and Projects in a CV Recommender System
We describe a CV recommender system built for the purpose of connecting candidates with projects that are relevant to their skills. Each candidate and each project is described by a textual document (CV or a project description) from which we extract a set of skills and convert this set to a numeric representation using two known models: Latent Semantic Indexing (LSI) and Global Vectors for Word Representation (GloVe) model. Indexes built from these representations enable fast search of similar entities for a given candidate/project and the empirical results demonstrate that the obtained l2 distances correlate with the number of common skills and Jaccard similarity.
LSI ; GloVe ; Recommender systems ; Nearest neighbors
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
nije evidentirano
Podaci o prilogu
17-29.
2020.
objavljeno
10.1007/978-3-030-59605-7_2
Podaci o matičnoj publikaciji
Artificial Intelligence and Mobile Services – AIMS 2020
Podaci o skupu
9th International Conference on Artificial Intelligence and Mobile Services – AIMS 2020
predavanje
18.09.2020-20.09.2020
Honolulu (HI), Sjedinjene Američke Države