A DECISION TREE BASED RECOMMENDATION SYSTEM FOR TOURISTS

Abstract

Choosing a tourist destination from the 
information that is available on the 
Internet and through other sources is one 
of the most complex tasks for tourists 
when planning travel, both before and 
during 
travel. 
Previous 
Travel 
Recommendation Systems (TRSs) have 
attempted to solve this problem. However, 
some of the technical aspects such as 
system accuracy and the practical aspects 
such as usability and satisfaction have 
been neglected. To address this issue, it 
requires a full understanding of the 
tourists' 
decision-making 
and 
novel 
models for their information search 
process. This paper proposes a novel 
human-centric TRS that recommends 
destinations to tourists in an unfamiliar 
city. It considers both technical and 
practical aspects using a real world data 
set we collected. The system is developed 
using a two-steps feature selection method 
to reduce number of inputs to the system 
and recommendations are provided by 
decision tree C4.5. The experimental 
results show that the proposed TRS can 
provide personalized recommendation on 
tourist destinations that satisfy the tourists.