The travel industry has been influenced by technological developments and has changed and grown day by day. Personalizing and improving user experiences has become impor- tant for this sector, which appeals to more users every day. This study aims to evaluate a system called "The Hybrid Offer Ranker" (THOR) in order to test toits performance to understand user preferences in the travel industry and provide personalized travel rec- ommendations based on these preferences. THOR is designed as a hybrid, personalized recommendation system to provide users with travel recommendations based on their past preferences. The main purpose of this study is to test the classification performance of the THOR system on users by creating a realistic data set. While creating this data set, users with different profiles were created. The route information created for each user was taken from the Google Maps API. A value was assigned to whether users received each offer or not, and the performance of the system in predicting this value was ex- amined. As a result, this study is an important one in terms of testing THOR, which was created to increase the effectiveness of recommendation systems in the travel industry and improve user experience based on realistic data to measure whether the system correctly understands user preferences and is successful in providing personalized travel recommendations.
Negli ultimi anni il settore dei viaggi ha visto un notevole sviluppo tecnologico e un cres- cente impatto delle applicazioni informatiche. Personalizzare e migliorare l’esperienza dell’utente è diventato importante per questo settore, che attira ogni giorno sempre più utenti. Questo studio mira a valutare un sistema chiamato "The Hybrid Offer Ranker" (THOR) al fine di testarne le prestazioni per comprendere le preferenze degli utenti nel settore dei viaggi e fornire consigli di viaggio personalizzati basati su queste preferenze. THOR è progettato come un sistema di consigli ibrido e personalizzato per fornire agli utenti consigli di viaggio basati sulle loro preferenze passate. Lo scopo principale di questo studio è testare le prestazioni di classificazione del sistema THOR sugli utenti creando un set di dati realistico. Durante la creazione di questo set di dati, sono stati creati utenti con profili diversi. Le informazioni sul percorso create per ciascun utente sono state prese dall’API di Google Maps. È stato assegnato un valore al fatto che gli utenti abbiano rice- vuto o meno ciascuna offerta ed è stata esaminata la prestazione del sistema nel prevedere questo valore. Di conseguenza, questo studio è uno studio importante in termini di test THOR, che è stato creato per aumentare l’efficacia dei sistemi di raccomandazione nel settore dei viaggi e migliorare l’esperienza dell’utente, sulla base di dati realistici, per misurare se il sistema comprende correttamente le preferenze dell’utente e ha successo nel fornire consigli di viaggio personalizzati.
Creation of a realistic dataset for the evaluation of a travel offer recommender system
Yilmaz, Melis
2023/2024
Abstract
The travel industry has been influenced by technological developments and has changed and grown day by day. Personalizing and improving user experiences has become impor- tant for this sector, which appeals to more users every day. This study aims to evaluate a system called "The Hybrid Offer Ranker" (THOR) in order to test toits performance to understand user preferences in the travel industry and provide personalized travel rec- ommendations based on these preferences. THOR is designed as a hybrid, personalized recommendation system to provide users with travel recommendations based on their past preferences. The main purpose of this study is to test the classification performance of the THOR system on users by creating a realistic data set. While creating this data set, users with different profiles were created. The route information created for each user was taken from the Google Maps API. A value was assigned to whether users received each offer or not, and the performance of the system in predicting this value was ex- amined. As a result, this study is an important one in terms of testing THOR, which was created to increase the effectiveness of recommendation systems in the travel industry and improve user experience based on realistic data to measure whether the system correctly understands user preferences and is successful in providing personalized travel recommendations.| File | Dimensione | Formato | |
|---|---|---|---|
|
2024_04_Yilmaz.pdf
accessibile in internet per tutti
Dimensione
1.03 MB
Formato
Adobe PDF
|
1.03 MB | Adobe PDF | Visualizza/Apri |
I documenti in POLITesi sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/10589/218183