Challenges in Recommender Systems for Tourism Manoj Reddy Dareddy University of California, Los Angeles California, USA mdareddy@cs.ucla.edu ABSTRACT Location Time Constraints In this position paper, we outline some of the challenges facing recommender systems in the tourism domain. The problems in this domain are unique compared to the traditional recommender systems. The challenges outlined in this paper include: dynamic itinerary planning, mobile platform, evaluation methods, group Recommender recommendation, social network, integration, serendipity, user System modeling, privacy and robustness. We provide an overview for each of the topics and present the opportunities for improvement. The tourism domain consists of a large amount of information stored digitally and recommender systems can act as a filter that Adaptive can personalize the experience for every tourist. Itinerary Categories and Subject Descriptors Figure 1: Itinerary generation H.4 [Information Systems Applications]: Miscellaneous; An important aspect of such systems is the human interface since Keywords it ultimately determines the interaction with the user. In this Tourism; Recommender Systems; Position paper regard, the design needs to ensure a minimal amount of cognitive effort on the user’s part. 1. INTRODUCTION Tourism broadly refers to the movement of people who are 2.2 Mobile exploring new places. Globally, [1] it accounts for 10% of the The future of computing is mobile. Mobile plays a very important world’s GDP and it supports about 1 in 11 jobs around the globe. role in this domain since tourists are always on the move. Hence, It is one of the fastest growing sectors and many nations depend it is important for recommender systems to take advantage of on it as a major source of income. It can be classified into various contextual information such as location, time of day etc. These categories based on their primary motive such as medical, mobile devices also allow different types of interactions to be educational, artistic, sports tourism etc. This domain consists of captured such as emotion, whether the user is travelling alone or enormous amount of information stored digitally that is not being with a group etc. The location information allows the system to used to its maximum potential. Recommender systems have huge recommend events, places to see that are physically close the user. opportunity in improving the experience of the tourists. This position paper presents various technical challenges that have not yet been addressed by the recommender system community in the tourism domain. The goal of this position paper is to discuss the open problems in this area for researchers to work on. 2. CHALLENGES 2.1 Dynamic Itinerary Planning One of the main challenges in this domain is optimal itinerary planning for tourists. Tourists generally have an agenda in mind of different places to visit in a city or events to attend, restaurants to try etc. There exist systems that recommend places to visit based on user interest but they are all static in nature. They do not Figure 2: Mobile devices have a huge role to play in tourism take into account changes that take place in real-time. For Another important aspect of mobile that shall play an important example, if a tourist would like to visit Paris, the system should be role in the future is its ubiquitous nature. This will ensure that the able to dynamically figure out the opening times and recommend user gets access to the right information at the right time and right an itinerary. There have been attempts to model this as an location. Current systems such as Google Now, perform such optimization problem where the objective function is to maximize ubiquitous computation by leveraging information from various a user specific satisfaction metric subject to constraints such as sources to personalize the user experience. opening times, budget etc. An example of a user specific metric could be the number of places visited or cost etc. 2.3 Evaluation Methods Copyright held by the author(s). The current evaluation methods for recommender systems mostly consider explicit feedback. The most popular techniques being RecTour 2016 - Workshop on Recommenders in Tourism held in conjunction with used are Root Mean Squared Error (RMSE) and MAE (Mean the 10th ACM Conference on Recommender Systems (RecSys), September 15, 2016, Boston, MA, USA. Average Error), which relies on explicit user feedback. challenge is to understand the preference of each user and filter out relevant information such as hotel deals etc. Advertising can play a very important role in recommender systems. A prime example of this is the Google Adwords RMSE error calculation program. It aims to provide relevant ads that are useful and the user is most likely to click. Similarly, for recommender systems, Another commonly used evaluation technique is Mean Average Error (MAE) which is defined as follows: ads play a very important role since they allow users to learn about relevant promotions such as hotel rooms, restaurant deals etc. Such an interface will allow a tourist to perform all relevant computation without having to switch between different applications which can be cumbersome. Also, it would be helpful Mean Average Error calculation if adequate information is provided for various places-of-interest. Both these metrics measure the difference between the predicted that the user is likely to visit. and actual value on a test dataset. These metrics depend on the explicit user information such as ratings feedback. Figure 2: Different forms of feedback by a tourist Recommender systems in the tourism domain need to be able to Figure 3: Integrated console of all tourist related activity gauge user satisfaction level by measuring their emotion in a minimally-intrusive manner. Some of the possible methods 2.7 Serendipity includes analyzing the user’s social media, pictures being taken Serendipity refers to the idea of discovering a new interest that the and explicit user feedback such as ratings or like/dislike. user had no idea about. These types of recommendations are the most effective but also the riskiest. The reward is high but the 2.4 Group Recommendation accuracy also tends to be low. In the tourism domain, if a user is Tourists generally travel in groups and current recommendation interested in art history, the user might be interested in ancient systems mainly focus on a single user rather than a group. The monuments which is a completely different interest. Such models main challenge is to combine individual preferences of different can be learnt using machine learning techniques that process large members and recommend items that are enjoyed by the group as a amounts of behavioral data. whole. Certain groups might be more interested in adventure activities whereas others might be inclined towards 2.8 User Modeling historical/cultural places. Some of the variables that need to There also needs to be better user modeling that is able to considered are: number of members in the group, individual understand latent user interests. In the tourism domain, the user restrictions and group characteristics. interests can be organized based on a taxonomy for example: nature, food, etc. This requires building new algorithms that can 2.5 Social Network scale better with different types of input data. Existing techniques Social connections play an important role in the recommendation such as collaborative filtering, matrix factorization etc. could be for tourism. For example, if a user’s friends recommend trying a applied in this area. Moreover, collaboration with tourism domain restaurant in a different city, then the user is likely to visit that experts shall help in better modeling of the user. restaurant. There are various types of social influence that ranges from different degrees. One possibility is to integrate existing 2.9 Privacy social network information from sites such as Facebook, Twitter Privacy plays a very important role in recommender systems. etc. The level of influence depends on the closeness of the user Since these systems have a lot of personal information, it becomes with another user, since it is more natural to trust close friends imperative to protect the privacy of the users. Current systems than users who are 3 or 4 degrees away. focus on differential privacy and use aggregates that prevent from identifying individual records. 2.6 Integration The main challenge facing tourists is the integration of various 2.10 Robustness sources of information. For example, the user needs to decide on The systems are vulnerable to manipulation and it becomes the airline, hotel, transportation method, tickets to various events important to protect them from various types of attacks. For etc. It is would be nice to have an end-to-end system that example, a malicious user might target a competitor by creating integrates such information in a condense format. The main fake accounts and down-rating their system, meanwhile increasing 4. ACKNOWLEDGMENTS the rating of own system. Our thanks to ACM SIGCHI for allowing us to modify templates they had developed. 3. CONCLUSION The tourism sector presents a number of opportunities for 5. REFERENCES recommender systems. There are many challenges some of which [1] World Tourism Organization UNWTO. (n.d.). Retrieved have been outlined in this position paper. These tourism domain June 23, 2016, from http://www2.unwto.org/content/why- specific problems require innovative approaches for implementing tourism recommender systems that can be used by a large number of tourists.