Volume XLII-4/W18
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLII-4/W18, 879–884, 2019
https://doi.org/10.5194/isprs-archives-XLII-4-W18-879-2019
© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.
Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci., XLII-4/W18, 879–884, 2019
https://doi.org/10.5194/isprs-archives-XLII-4-W18-879-2019
© Author(s) 2019. This work is distributed under
the Creative Commons Attribution 4.0 License.

  19 Oct 2019

19 Oct 2019

A MULTI CRITERIA RECOMMENDATION MODEL FOR JAUNT

A. Razavi and F. Hosseinali A. Razavi and F. Hosseinali
  • Department of Surveying Engineering, Faculty of Civil Engineering, Shahid Rajaee Teacher Training University, Tehran, Iran

Keywords: Recommender Systems, Logistic Regression, Content Based Filtering, Classification, Itinerary, Points of Interest

Abstract. Nowadays, people in most parts of the world always visit, travel and have fun in their cities or other cities, and they spend considerable time and money in their city or in other cities as a tourist. The existence of an intelligent and automated system that can provide the most suitable recreational and cultural offerings at any time and place, with regard to financial capability and time and transport constraints, as well as individual interests and personalization; has always been felt. Recommender systems can be used to suggest suitable recreational options for the user. The main difference between the recommendation model in this study and the previous models is to focus on the short-term planning of a few hours for one day. Previous models were often based on planning a few days a week or days of the month. Also, the cost factor has been considered in this research, which has been less considered in previous models. We used collaborative filtering based on logistic regression to predict whether a type of places is a proper proposition to a user or not. Our case study is about recommending the board game cafés in the city of Kerman, Iran and the result shows that mixed groups between 15 to 30 years old are the best target and our model can predict if board game café is a good suggestion to different users. We used correlation based recommender systems when board game cafes are a proper suggestion for a user and there are at least two options for the user. In case there is no information about the user and his previous rating, popularity based recommender system can be useful. We also used content based recommender systems to give recommendations by having some background information about previous itineraries of a user and his rating to those.