TourWithMe: Recommending peers to visit atractions together

Abstract

When a user travels alone or in a small group, usually likes to share the experience of visiting diferent attractions in a larger group. This article propose TourWithMe, our first approach to the problem of recommending peers to visit attractions in a city together. To this aim, TourWithMe automatically learns the user's interests from previously visited attractions, that are then combined with explicit preferences provided by the user to find compatible tourists in the city. TourWithMe recommends to the user diferent groups and, for each group, attractions that they would enjoy visiting together.

CCS CONCEPTS

• Information systems → Recommender systems; Social recommendation; Crowdsourcing. group recommender system; tourism; crowdsourcing; user modeling

INTRODUCTION

Visiting a new city is always a challenging experience. Among the set of touristic attractions available in the city, tourists have to select, and usually prioritize, those that are more appealing according to their interests, available time and budget. In consequence, planning a holiday is usually a stressful activity and travellers relay in the use of diferent applications that may support their decision-making processes.

Recommender systems for tourism arisen to cope with the information overload to which tourists face when visiting a new city. In this regard, recommender systems have focused on diferent aspects of the domain, such as recommending hotels [ 1, 25 ], routes [ 10, 16 ], restaurants [ 9 ], itineraries [ 7, 15 ], and attractions [ 13, 33, 34 ].

A hot topic in recommender systems research is the recommendation of items to groups of users, since recommendations need to satisfy a group of users as a whole, instead of individual users [ 5, 6 ]. In the field of tourism, recommender systems for groups have been proposed for users who travel with a predefined group (for example, a group of friends or family travelling together) [ 2, 11 ].

To the best of our knowledge, none of the existing approaches considered the proposal of groups to visit diferent attractions together. This kind of recommender system might be extremely useful for users who visit a destination alone or in a small group (for example, with his/her couple) and who want to meet peers to share the experience of touring together. The need of this kind of service becomes clear given the existence of many websites 1,2,3 and social network groups 4,5,6 dedicated to people who wants to meet other people and form groups for tourism.

In this context, the popularization of mobile devices brings forward new challenges and opportunities for the implementation of personalized applications and location-aware services. Particularly, mobile devices enable to capture the user’s mobility history and taking advantage of geographic proximity of other users to enhance the user experience [ 14 ].

In this article, we present TourWithMe, a recommender system in the tourism domain that takes advantage of mobile devices for recommending travellers to form groups to visit attractions or points of interest (POI) together. Our approach considers geolocalization provided by mobile devices in two ways. On the one hand, the approach implicitly learns the user’s interest from the places he/she visits, the amount of time spent in each place, and the time spent travelling to those places. In this way, users do not have to manually check-in every place they visit or to explicitly provide their interests, as required by most of the current approaches. On the other hand, the approach finds other tourists in the proximity of the user and suggests forming a group with those users who have similar interests. Once a group is formed, the approach suggests to visit nearby venues that the whole group would enjoy visiting.

The remainder of this paper is organized as follows. Section 2 discusses related works about recommenders system for tourism. Section 3 presents the proposed approach for recommending travellers forming groups to visit attractions together. Finally, Section 4 presents conclusions and future works. 2

RELATED WORK

Recommenders System for tourism is a hot topic that has been addressed in several works in the last years. These works proposed approaches to recommend users to visit a nearby POI or even a tour itinerary. To carry out this task, proposed approaches used diferent information, such as the user’s current location, information about nearby POIs, user preferences and interests, current day and time, temporal restrictions, etc. The kind of information used and the way in which this information is obtained vary depending on the approach.

In [ 31 ] and [ 19 ], authors asked users to manually provide their interest and preferences. Both approaches recommend a personalized tour itinerary that fits the user’s interests. To carry out this

1https://www.yourtravelmates.com/

2https://www.workaway.info/ 3https://www.couchsurfing.com/ 4https://www.facebook.com/groups/altmtl/ 5https://www.facebook.com/groups/1157818554266712/ 6https://www.facebook.com/groups/travellinks/ task, [ 31 ] used a Greedy algorithm while [ 19 ] used an evolutionary algorithm. The main disadvantages of these works are that manually introducing interests may be a stressful task for users, and they tend to be reluctant to explicitly provide this kind of information [ 26 ]. For this reason, some works in the literature proposed to automatically infer the user’s interests by analyzing the previously visited POIs.

To address this task, some works used check-ins made by user in location based social networks (LBSN) [ 17, 35 ] and geotagged photos from social networks [ 4, 20, 21 ] in order to reconstruct the history of visited POIs. In [ 4, 17, 20 ], authors proposed approaches that infer the interests of the user for each POI category according to the number of visited POIs belonging to that category. These approaches use these interests to generate a ranking of possible POIs to be visited by the user. In [ 35 ], authors proposed a similar approach that infers the user’s interest from Jiepang check-ins data. As the user’s interests may change according to the time of day, this approach also divides the day into six time slots and calculates the user’s interests for each time slot separately. In [ 21 ], authors proposed an approach that calculates the duration of each visit by considering the timestamps of the first and the last photos took in the visited POI. The approach uses this information to estimate the user interest for a POI category. For example, if the user spends more time in museums than the average time spent by other users, the approach infers that the user is interested in museums.

As some tourists tend to travel in group, recommending POIs to a group of users instead of to a single user is a useful feature in the tourism domain. Some approaches in the literature address this feature by combining the users’ profiles into a single group profile [ 12, 27 ]. In this way, approaches designed for recommending POIs to a single profile (usually a user profile) can recommend also POIs to a group by taking the group profile as input. There are two main approaches to combine user profiles: aggregation, when the resultant group profile is the union of all the group members preferences; and intersection, when the resultant group profile is the intersection of all the group members preferences. The approach presented by [ 5 ] used an hybrid approach for generating recommendations to groups of tourists, which combines the demographic information of users, the ratings of the community and the content-specific information about the items. The individual ratings inferred from the hybrid profile are weighted according to a fixed set of social relationships among the members of the group. Finally, the influenced individual ratings of all members of a group are combined to estimate a group rating for diferent items.

To the best of our knowledge, none of the existing approaches considered the proposal of groups to visit diferent attractions together. The most similar approach to the one presented in this article is the one presented in [ 22 ]. In this work, authors proposed an approach oriented to assisting travel agencies for grouping tourists. The approach uses K-means algorithm to cluster a predefined set of users into K groups. Each resultant group contains users with similar interest. Then, the approach assigns a tour itinerary from a set of predefined tour itineraries to each group of users. However, this approach is not useful for a tourist who is alone in an unknown city and wants to meet peers to visit POIs together. 3

SYSTEM DESIGN

This step consists of analyzing the mobility data of the user in order to infer his/her preferences. In order to carry out this task, TourWithMe takes advantage of modern mobile devices. These devices are equipped with several sensors that allow estimating the location of the user. For example, it is possible to estimate the user location by knowing the nearby WiFis or by using the GPS of the smartphone. By tracking the user location, TourWithMe detects visits to places, also named stay points. A stay point is defined in the literature as a geographic region where the user stayed over a time threshold Ts within a distance threshold Ds [ 24, 29, 32 ]. In particular, TourWithMe detects a visit when the user stays for more than 5 minutes within a distance of 50 meters. Each visit is represented as a tuple (C, Ti , Te ), where C is the centroid of the geographic area where the user stayed, Ti is the start time of the visit and Tf is the end time of the visit.

When a visit is detected, TourWithMe identifies the POI visited by the user, if any. To carry out this task, TourWithMe relies on public data extracted from OpenStreetMap7 (OSM). In particular, TourWithMe uses the Overpass Turbo API8 to query POIs that are less than 50 meters away from the visit. If there are no nearby POI, it is considered that the user stayed in some other place (e.g. in a store). If there is more than one nearby POI, TourWithMe selects the POI with the highest score according to Equation 1. This equation compares the duration of a visit V of user U and the average time of visit for a POI P . The average time of visit for P is computed from previous visits of other users to the same POI. It is important to notice that the user can manually modify the visited POI if needed. score(V , P ) = 1 − |avдDurationO f V isit (P ) − duration(V )| avдDurationO f V isit (P ) (1)

Once the visit has an associated POI, TourWithMe estimates the interest of the user in that POI. The interest of the user in a POI is a real value between 0 and 1 where 0 means that the user is not interested in the POI and 1 corresponds to the maximum interest. This value is computed according to Equation 2 and considers the

7https://www.openstreetmap.org/ 8http://overpass-turbo.eu/

time that the user spent in the POI (intvisit −t ime ) and the time of the travel T to that POI (intt r avel −t ime ).

int (T , V , P ) = intvisit −t ime (V , P ) + intt r avel −t ime (T , P ) (2) 2

To compute the first term of the equation, in [ 21 ] authors proposed to compute the ratio between the time spent by the user in the POI and the average duration of visits to that POI. However, this approach is not useful when a POI has diferent groups of users who visit the POI with diferent average times. For example, a museum can ofer 1-hour and 2-hours guided tours. An average of 1.5 hours is then not representative for a user taking the 1-hour tour nor to a user taking the 2-hours tour. Furthermore, computing the interest of a user in a POI in this way doesn’t give a normalized value of the user interest. To overcome the above-mentioned problems, TourWithMe uses the cumulative percentage of duration of visits. Equation 3 shows how the approach computes intvisit −t ime (V , P ) for a visit V to a POI P . For example, if spent 14 minutes in P , and 60% of people stayed less than 14 minutes in P , then the interest of the user in P is 0.6.

Pdur at ion(V ) Vd,p

d=0 intvisit −t ime (V , P ) =

Vp where Vd,p is number of visits to POI p with a duration d, and Vp is the number of visits to POI p.

The second term of Equation 2, intt r avel −t ime (T , P ), compares the time spent by a user in a POI with respect to the time spent travelling to that POI. In [ 8 ] authors proposed travel-time ratio as a way to calculate how much time a user is willing to travel to perform an activity. In [ 30 ], authors found higher travel-time ratios for activities in which users are interested, such as sport and recreation activities. Mapping the conclusions arrived in the above-mentioned research to the tourism domain, we can assume that if a user travels a long time to visit a given POI, he/she has a great interest in that POI. Equation 4 details how to calculate this ratio for simple journeys in which the user goes to a POI and then (3) returns to his place of lodging. The way to calculate the time ratio for journeys in which the user visit several POIs before returning his/her place of lodging is detailed in [ 30 ].

duration(T ) intt r avel −t ime (T , V ) = duration(T ) + duration(V ) (4) By knowing the interest of the user in each POI he/she visited, it is possible to estimate his/her interest for each POI category. As POIs are extracted from OSM, they have diferent pairs of key-value describing them. For example, {”tourism” : ”museum”}, {”name” : ”Le Louvre”}. These pairs of key-value are used to label the POI with POI categories. For example, "Le Louvre" is categorized as a "museum". To calculate the interest of a user for a specified POI category C, TourWithMe calculates the average interest of the user in every POI p belonging to C that he/she previously visited (Equation 5).

intinf er r ed (U , C) =

Pp ∈C interest (U , p) |C | (5) 3.2

Forming groups

For suggesting groups to a user, TourWithMe considers three factors: geolocalization, user’s preferences and similarity between users’ interests regarding POIs categories. When the user asks for suggestions or when he/she arrives to a new city, TourWithMe first ifnd the set of users SR within a parameter radio R from the user’s current location. If R is not set by the user, TourWithMe considers the set of users visiting the same city. SR contains then the set of candidate users near to the user’s location.

Once the set of candidate users is obtained, it is filtered by the user’s preferences. User’s preferences are a list of restrictions that the user is able to manually fill in his/her profile, and indicate the system what kind of users are expected to be recommended to the target user. These preferences, which are all optional, include: • age range: indicates the minimum and maximum age of other users in the group • sex: preferred sex of people in the group (male, female, any) • languages: a list of languages that users in the group should speak • country of residence: if the user prefers other users from specific countries • children preference: users can indicate whether they prefer tourists traveling with children or not.

Then, if a user established in his/her profile that he/she prefers other tourists aging between 20 and 30, any candidate whose age is outside those limits is removed from the set of candidates. The resulting set Sf contains the set of compatible candidates with the user’s preferences.

Other kind of preferences included in the user profile are the following: • a list of categories of interest: an explicit list of the POI categories in which the user manually indicated interest. Categories are taken from the OpenStreetMap Semantic Network [ 3 ]. • budget: indicates the amount of money the user expects to spend while visiting attractions. This variable is discretized in four values (0, $, $$, $$$), indicating free, cheap, moderate, and expensive POIs, respectively

The list of categories manually defined by the user and the inferred interests (which were obtained as described in Section 3.1) are combined to define the real interest of a user U in a category C (Equation 6). If user U explicitly indicated interest in C (by adding it to his/her list of interests), then int (U , C) is the average between 1 and intinf er r ed (U , C). Otherwise, int (U , C) is equals to intinf er r ed (U , C).

 1+intinf e2r r ed (U ,C) , if U is interest in C int (U , C) =  (6) intinf er r ed (U , C), otherwise 

In the current implementation of TourWithMe, each candidate user v in Sf is ranked by computing the soft cosine similarity with respect to the target user U (Equation 7). This similarity measure does not assume that features in the space model are independent and then introduce the similarity of features into the equation of the traditional cosine similarity.

PiN, j si jUivj

so f t _cosine(U , v) = qPiN, j si jUiUj qPiN, j si j civj (7) where Ui is the ith feature for user U , vi is the ith feature for user v, and si j is the similarity between the ith and the jth features. The similarity between features i and j, si j , is computed by using the semantic similarity of OSM tags [ 3 ]. The set SC ⊂ Sf with the K most similar users is considered for forming groups in the next step.

When a user U asks for a group recommendation, he/she must define a preferred group size Z (where Z < K ). Then, from Sc , all possible groups of size Z including the target user U are computed, and a cohesion score is assigned to each of them. Cohesion is computed as the average similarity between each pair of users in the group. Groups are finally sorted by the cohesion score. 3.3

Recommending POIs

Although groups are formed by finding tourists with similar interests, diferent users always will have some diferent interests. To address these diverse interests, most approaches in the literature build a group interest profile by aggregating or by intersecting the preferences of all group members [ 11, 13, 27, 28 ]. From these two options, aggregating preferences is preferable since it allows introducing serendipity in the recommendations enabling the user to discover attractions that may not be recommended by a recommender system for individuals. Serendipitous items are items that users would not find by themselves or even look for, but that would enjoy consuming. The introduction of serendipity in recommender systems is fundamental to avoid users losing the interest in recommendations due to a overspecialization of the system in the user’s already-known interests [ 18 ]. This overspecialization, avoids the recommender system to learn new interests of the user, and enables the user to be able to predict by themselves what items would be recommended by the system, reducing in consequence the user’s satisfaction with the recommendations.

For example, Figure 2 shows a group of three users with their respective interests. By aggregating user interests, the interest of the resultant group profile in a category Ci is the average interest of the three users in Ci . In the literature, this is known as average aggregating strategy [ 23 ]. As the interest of user B in C2 is not defined, the interest of the whole group in C2 is calculated by considering only users A and C. Thus, the resultant group profile has a high interest in category C2. In this way, if the approach recommends a POI of C2, it will encourage User B to visit a new kind of POI. Instead, by intersecting user interests, the resultant group profile will not have any interest value defined for C2, since not all users of the group have an interest defined in C2. Thus, the approach will encourage users to continue visiting the same kind of POIs they already visited before. TourWithMe builds a group interest profile based on the average interest preference of all group members. Given a group д = u1, ..., uk , the group interest in a cagetory c is defined according to Equation 8. where Cp are the categories associated to POI p.

Continuing with the example of Figure 2, by using the average interest of all the group members not necessary may lead to making the best recommendation. For example, Figure 2 shows that the group profile has an interest of 0.67 for C1 and 0.65 for C3. Thus, recommending a POI belonging to C1 would be preferred than recommending a POI belonging to C3. However, the variation of interests for this category is very high: User A has an interest of 0.81, while User B has an interest of 0.55. Thus, visiting a POI belonging to C1 seems to be unfair for User B. Moreover, User A will want to stay in the POI a longer time, while user B will want to leave before. Instead, when visiting a POI belonging to C3, the three users will have a similar interest in the POI and there are more chances that they will agree about how long to be in that place.

To considering this situation, TourWithMe looks for recommending the POI that best fits the group profile at the same time that it reduces the variation of interest among users for the recommended POI. Equation 10 shows how TourWithMe score a POI p for a group д. All POIs in the user’s neighbourhood are ranked according to this equation, and the top-5 POIs are assigned to each group as recommendations.

max Inter est (д, p) − minInter est (д, p) scor e(д, p) = int (д, p)− ation of interest between the members of group д for POI p. |д| where max Inter est (д, p)−minInter est (д, p) is the maximum variAlong to each recommendation TourWithMe computes the estimated time that the group would spend in each POI by using the cumulative percentage of duration of visits, as detailed in Equation 3. In this case, if the group interest in the category of a POI p is, for example, 0.6 and the time spent by 60% of the people at the given POI is t , we assign t as the estimated time that the group would (10) spend at p. 4

CONCLUSIONS AND FUTURE WORK

In this article we presented TourWithMe, a first approach to the problem of recommending peers to visit diferent attractions in a group. We believe that our approach might appeal tourists traveling alone or in small groups to enhance the experience of enjoying the attractions ofered by a new city.

TourWithMe is currently in a prototype stage, and is developed as a native Android application. This application tracks user location and detect visits when the user stays for more than 5 minutes within a distance of 50 meters. Then, TourWithMe associates each visit to a POI extracted from OpenStreetMap when possible. In addition, TourWithMe identifies the transport mode of each travel, which in the future may be a useful feature for POI recommendation. For example, if user moves by car, it is possible to recommend more distant POIs than if he/she moves on foot.

The next step in our research is to evaluate our approach with a benchmark dataset. As there is no benchmark dataset available for POI recommendation for group of users, most works in the literature use datasets with individual ratings and simulate groups. The rating of a simulated group for a POI may be estimated as the average ratings of the group members. The main challenge after evaluating the proposed approach with a simulated dataset will naturally be the validation with real users.

References

[1] Marie Al-Ghossein , Talel Abdessalem , and Anthony Barré . 2018 . Cross-Domain Recommendation in the Hotel Sector . In Proceedings of RecTour 2018 , Julia Neidhardt , Wolfgang Wörndl, Tsvi Kuflik, and Markus Zanker (Eds.). 1 - 6 .

[2] Aris Anagnostopoulos , Reem Atassi, Luca Becchetti, Adriano Fazzone, and Fabrizio Silvestri . 2016 . Tour recommendation for groups . Data Mining and Knowledge Discovery 31 , 5 (sep 2016 ), 1157 - 1188 . https://doi.org/10.1007/s10618-016- 0477-7

[3] Andrea Ballatore , Michela Bertolotto, and David C. Wilson . 2012 . Geographic knowledge extraction and semantic similarity in OpenStreetMap . Knowledge and Information Systems 37 , 1 (oct 2012 ), 61 - 81 . https://doi.org/10.1007/s10115-012- 0571-0

[4] Igo Ramalho Brilhante , Jose Antonio Macedo, Franco Maria Nardini, Rafaele Perego, and Chiara Renso . 2015 . On planning sightseeing tours with TripBuilder . Information Processing & Management 51 , 2 (mar 2015 ), 1 - 15 . https://doi.org/10. 1016/j.ipm. 2014 . 10 .003

[5] Ingrid Christensen , Silvia Schiafino, and Marcelo Armentano . 2016 . Social group recommendation in the tourism domain . Journal of Intelligent Information Systems 47 , 2 (mar 2016 ), 209 - 231 . https://doi.org/10.1007/s10844-016-0400-0

[6] Amra Delic , Julia Neidhardt, Thuy Ngoc Nguyen, and Francesco Ricci . 2018 . An observational user study for group recommender systems in the tourism domain . Information Technology & Tourism 19 , 1 - 4 (feb 2018 ), 87 - 116 . https: //doi.org/10.1007/s40558-018-0106-y

[7] Linus W Dietz and Achim Weimert . 2018 . Recommending Crowdsourced Trips on wOndary . In Proceedings of RecTour 2018 , Julia Neidhardt , Wolfgang Wörndl, Tsvi Kuflik, and Markus Zanker (Eds.). 13 - 17 .

[8] Martin Dijst and Velibor Vidakovic . 2000 . Travel time ratio: the key factor of spatial reach . Transportation 27 , 2 ( 2000 ), 179 - 199 . https://doi.org/10.1023/a: 1005293330869

[9] Haoxian Feng and Thomas Tran . 2018 . Context-Aware Approach for Restaurant Recommender Systems . In Encyclopedia of Information Science and Technology, Fourth Edition. IGI Global , 1757 - 1771 . https://doi.org/10.4018/978-1- 5225 -2255- 3. ch153

[10] Ander Garcia , Maria Teresa Linaza, Olatz Arbelaitz, and Pieter Vansteenwegen . 2009 . Intelligent Routing System for a Personalised Electronic Tourist Guide . In Information and Communication Technologies in Tourism 2009 . Springer Vienna, 185 - 197 . https://doi.org/10.1007/978-3- 211 -93971-0_ 16

[11] Inma Garcia , Laura Sebastia, and Eva Onaindia . 2011 . On the design of individual and group recommender systems for tourism . Expert Systems with Applications 38 , 6 (jun 2011 ), 7683 - 7692 . https://doi.org/10.1016/j.eswa. 2010 . 12 .143

[12] Inma Garcia , Laura Sebastia, Eva Onaindia, and Cesar Guzman . 2009 . A Group Recommender System for Tourist Activities . In E-Commerce and Web Technologies . Springer Berlin Heidelberg, 26 - 37 . https://doi.org/10.1007/978-3- 642 -03964- 5 _ 4

[13] Damianos Gavalas and Michael Kenteris . 2011 . A web-based pervasive recommendation system for mobile tourist guides . Personal and Ubiquitous Computing 15 , 7 (may 2011 ), 759 - 770 . https://doi.org/10.1007/s00779-011-0389-x

[14] Damianos Gavalas , Charalampos Konstantopoulos, Konstantinos Mastakas, and Grammati Pantziou . 2014 . Mobile recommender systems in tourism . Journal of Network and Computer Applications 39 (mar 2014 ), 319 - 333 . https://doi.org/10. 1016/j.jnca. 2013 . 04 .006

[15] Damianos Gavalas , Charalampos Konstantopoulos, Konstantinos Mastakas, and Grammati Pantziou . 2014 . A survey on algorithmic approaches for solving tourist trip design problems . Journal of Heuristics 20 , 3 (mar 2014 ), 291 - 328 . https://doi.org/10.1007/s10732-014-9242-5

[16] Koji Kawamata and Kenta Oku . 2019 . Roadscape-based Route Recommender System Using Coarse-to-fine Route Search . Journal of Information Processing 27 , 0 ( 2019 ), 392 - 403 . https://doi.org/10.2197/ipsjjip.27.392

[17] K. Kesorn , W. Juraphanthong , and A. Salaiwarakul . 2017 . Personalized Attraction Recommendation System for Tourists Through Check-In Data . IEEE Access 5 ( 2017 ), 26703 - 26721 . https://doi.org/10.1109/access. 2017 .2778293

[18] Denis Kotkov , Jari Veijalainen, and Shuaiqiang Wang . 2016 . Challenges of Serendipity in Recommender Systems . In Proceedings of the 12th International Conference on Web Information Systems and Technologies. SCITEPRESS - Science and and Technology Publications . https://doi.org/10.5220/0005879802510256

[19] Yohei Kurata and Tatsunori Hara . 2013 . CT-Planner4: Toward a More UserFriendly Interactive Day-Tour Planner . In Information and Communication Technologies in Tourism 2014 . Springer International Publishing, 73 - 86 . https: //doi.org/10.1007/978-3- 319 -03973- 2 _ 6

[20] Kwan Hui Lim . 2015 . Recommending Tours and Places-of-Interest based on User Interests from Geo-tagged Photos . In Proceedings of the 2015 ACM SIGMOD on PhD Symposium - SIGMOD '15 PhD Symposium . ACM Press. https://doi.org/10. 1145/2744680.2744693

[21] Kwan Hui Lim , Jefrey Chan, Christopher Leckie , and Shanika Karunasekera . 2017 . Personalized trip recommendation for tourists based on user interests, points of interest visit durations and visit recency . Knowledge and Information Systems 54 , 2 (may 2017 ), 375 - 406 . https://doi.org/10.1007/s10115-017-1056-y

[22] Kwan Hui Lim , Jefrey Chan, Christopher Leckie , and S hanika Karunasekera. 2016 . Towards next generation touring: Personalized group tours . In Twenty-Sixth International Conference on Automated Planning and Scheduling . 421 - 430 .

[23] Judith Masthof . 2015 . Group Recommender Systems: Aggregation, Satisfaction and Group Attributes. In Recommender Systems Handbook. Springer US , 743 - 776 . https://doi.org/10.1007/978-1- 4899 -7637-6_ 22

[24] Raul Montoliu and Daniel Gatica-Perez. 2010 . Discovering human places of interest from multimodal mobile phone data . In Proceedings of the 9th International Conference on Mobile and Ubiquitous Multimedia - MUM '10 . ACM Press. https: //doi.org/10.1145/1899475.1899487

[25] Qing Qi , Jian Cao, Yudong Tan , and Quanwu Xiao . 2018 . Cross-Domain Recommendation Method in Tourism . In 2018 IEEE International Conference on Progress in Informatics and Computing (PIC) . IEEE. https://doi.org/10.1109/pic. 2018 .8706265

[26] Feng Qiu and Junghoo Cho . 2006 . Automatic identification of user interest for personalized search . In Proceedings of the 15th international conference on World Wide Web - WWW '06 . ACM Press. https://doi.org/10.1145/1135777.1135883

[27] Logesh Ravi and Subramaniyaswamy Vairavasundaram . 2016 . A Collaborative Location Based Travel Recommendation System through Enhanced Rating Prediction for the Group of Users . Computational Intelligence and Neuroscience 2016 ( 2016 ), 1 - 28 . https://doi.org/10.1155/ 2016 /1291358

[28] Senjuti Basu Roy , Laks V.S. Lakshmanan , and Rui Liu. 2015 . From Group Recommendations to Group Formation. In Proceedings of the 2015 ACM SIGMOD International Conference on Management of Data - SIGMOD '15 . ACM Press. https://doi.org/10.1145/2723372.2749448

[29] Xiao Wen Ruan , Shou Chung Lee, and Wen Chih Peng . 2014 . Exploring LocationRelated Data on Smart Phones for Activity Inference . In 2014 IEEE 15th International Conference on Mobile Data Management . IEEE. https://doi.org/10.1109/ mdm. 2014 .71

[30] Yusak O. Susilo and Martin Dijst . 2010 . Behavioural decisions of travel-time ratios for work, maintenance and leisure activities in the Netherlands . Transportation Planning and Technology 33 , 1 (feb 2010 ), 19 - 34 . https://doi.org/10. 1080/03081060903429280

[31] Pieter Vansteenwegen , Wouter Soufriau, Greet Vanden Berghe, and Dirk Van Oudheusden. 2011 . The City Trip Planner: An expert system for tourists . Expert Systems with Applications 38 , 6 (jun 2011 ), 6540 - 6546 . https://doi.org/10.1016/j. eswa. 2010 . 11 .085

[32] Yang Ye , Yu Zheng, Yukun Chen, Jianhua Feng, and Xing Xie . 2009 . Mining Individual Life Pattern Based on Location History . In 2009 Tenth International Conference on Mobile Data Management: Systems, Services and Middleware . IEEE. https://doi.org/10.1109/mdm. 2009 .11

[33] Hongzhi Yin , Weiqing Wang, Hao Wang , Ling Chen , and Xiaofang Zhou . 2017 . Spatial-Aware Hierarchical Collaborative Deep Learning for POI Recommendation . IEEE Transactions on Knowledge and Data Engineering 29 , 11 (nov 2017 ), 2537 - 2551 . https://doi.org/10.1109/tkde. 2017 .2741484

[34] Hongzhi Yin , Xiaofang Zhou, Bin Cui, Hao Wang, Kai Zheng , and Quoc Viet Hung Nguyen. 2016 . Adapting to User Interest Drift for POI Recommendation . IEEE Transactions on Knowledge and Data Engineering 28 , 10 (oct 2016 ), 2566 - 2581 . https://doi.org/10.1109/tkde. 2016 .2580511

[35] Zhiwen Yu , Huang Xu, Zhe Yang , and Bin Guo . 2016 . Personalized Travel Package With Multi-Point-of-Interest Recommendation Based on Crowdsourced User Footprints . IEEE Transactions on Human-Machine Systems 46 , 1 (feb 2016 ), 151 - 158 . https://doi.org/10.1109/thms. 2015 .2446953