Orchestration of POIs Ubiquitous Contexts: a Review of Recommendation Systems Based on Matrix Factorisation Model

EasyChair Preprint no. 11607

Abstract

Currently, to take full advantage of the capabilities of Artificial Intelligence (AI), Smart tourism must use Context-Aware Recommendation Systems (CARS) to orchestrate the evolving contexts of users with smartphones in order to improve their travel experiences. This type of orchestration allows points of interest (POIs) recommendations to be personalised according to the ubiquitous context of tourists during their visits. Recommending the next POIs to visit can be based on collaborative filtering techniques founded on memory or models such as matrix factorisation (MF). This paper explains the contribution of approaches that integrate contexts into models, such as MF, compared to collaborative filtering approaches without context. Consequently, this survey shows that collaborative filtering techniques using MF considerably alleviate the problems associated with the cold start of CARS and that the three types of orchestration of tourist contexts (pre-filtering, post-filtering or context modelling) improve their satisfaction.

Keyphrases: cars, collaborative filtering, Context Modelling, context post-filtering, context pre-filtering, Matrix Factorisation, Orchestration, POI

@Booklet{EasyChair:11607,
  author = {Hadjhenni Mhamed and Dennouni Nassim and Slama Zohra},
  title = {Orchestration of POIs Ubiquitous Contexts: a Review of Recommendation Systems Based on Matrix Factorisation Model},
  howpublished = {EasyChair Preprint no. 11607},

  year = {EasyChair, 2023}}