An improved approach contextual suggestion system for E-Tourism
<p>The contextual suggestion system is defined as generating a list of venues for a user, based</p><p>on temporal and geographical context as well as travellers preferences relating to venues to</p><p>be suggested. A lack of effec...
Saved in:
| Main Author: | |
|---|---|
| Format: | thesis |
| Language: | eng |
| Published: |
2022
|
| Subjects: | |
| Online Access: | https://ir.upsi.edu.my/detailsg.php?det=9536 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | <p>The contextual suggestion system is defined as generating a list of venues for a user, based</p><p>on temporal and geographical context as well as travellers preferences relating to venues to</p><p>be suggested. A lack of effective methodologies has compromised the accuracy of the</p><p>contextual suggestion system in e-tourism. In this regard, the Text Retrieval Conference</p><p>(TREC) has been organized yearly to focus not only on the development of information</p><p>retrieval systems but also on the approaches leading to the improvement of the systems.</p><p>Besides, TREC provides datasets and standard protocols for evaluation to ensure fair</p><p>comparisons. In the study, an improved approach based on four main phases has been</p><p>proposed for the contextual suggestion system in e-tourism, namely, (i) Dataset Enrichment,</p><p>(ii) Profile Enrichment, (iii) User Modelling, and (iv) Ranking Suggestion. The TREC</p><p>dataset is used to evaluate the proposed approach. In the Dataset Enrichments improvement,</p><p>tags prediction, semantic similarity between tags, and correlation between tags are used. The</p><p>improvement in Profile Enrichment is based on context processing and relevancy between</p><p>the user and venue profiles in the given context. On the other hand, the improvement in User</p><p>Modelling is based on content-collaborative filtering and iterative-based approaches. Lastly,</p><p>a linear combination of true rocchio and cosine similarity is used to improve Ranking</p><p>Suggestion. The performance of the proposed approach is evaluated based on TRECs</p><p>standard evaluation protocols consisting of NDCG@5, P@5, and MRR. The experimental</p><p>results show an increment of 5% to 12% of accuracy in the proposed approach and the</p><p>increment is significantly better than the baseline run. In conclusion, the proposed approach</p><p>shows significant improvements consisting of 12.5% in P@5, 4.77% in NDCG@5, and</p><p>5.04% in MRR. This study implicates that the use of a contextual-based personalized venue</p><p>suggestions system enhances the travel experience of a traveller.</p> |
|---|