A Motivation-Aware Approach for Point of Interest Recommendations Khadija Ali Vakeel Sanjog Ray Indian Institute of Management Indore Indian Institute of Management Indore Prabandh Shikhar, Indore Prabandh Shikhar, Indore India India 091-0731-2439-666 091-0731-2439-524 f13khadijav@iimidr.ac.in sanjogr@iimidr.ac.in ABSTRACT mobile phones. Most existing context aware recommender systems primarily use Tourism industry is hugely impacted by the ubiquity of mobile a combination of ratings data, content data like features or phones in consumer lives [3]. Availability of many travel related attributes of the product or service, context data like location or apps and ease of access of free Wi-Fi spots has made mobile time and social network data. In this paper, we propose a novel phones the main decision making tool in helping tourists make approach for refining the recommendations made by location- travel related decisions. Mobiles phones complemented with aware recommender systems based on user motivations for intelligent travel related apps has completely transformed the checking in at locations in location based social networks. Based travel experience [4]. Among the technologies used for on a classification that classifies user’s motivation for checking in applications created for tourism, location aware and context aware at a Point Of Interest into seven categories we propose an based apps are the most popular as they have helped tourists to approach that will help refine recommendations in a way that can enhance their travel experience by making relevant be better explained to the user. We also show the applicability of recommendations. There is still a need for developing new our approach by analyzing a dataset extracted from Foursquare. approaches for recommending point of interests to tourists based on the variety of contextual and personal data available. This CCS Concepts paper tries to address the above need by proposing a novel • H.3.3 [Information Storage and Retrieval]: Information approach for recommending users places, restaurants, events etc. Search and Retrieval—Information Filtering based on user motivation profile that is derived from his check-in data from location based social networks. Keywords Location-based social networks; Point Of Interests In this paper, we propose a novel approach for refining the Recommendations; Motivation-Aware; Explanations. recommendations made by location-aware recommender systems based on user motivations. Most existing recommender systems 1. INTRODUCTION primarily use a combination of ratings data, content data like Availability of multiple product choices and easy access to features or attributes of the product or service, and context data information about them has made the task of making correct like location or time. We propose to integrate the user checking in purchase decision, by evaluating the information available, a huge motivation at places he has visited places into the location-aware problem for the consumers of the products or services. recommendation system, as it will help refine recommendations in Recommender systems are software that helps customers make a way that can be better explained to the user. This will also lead these decisions by providing them product recommendations that to increased adoption of the recommendations as prior research are relevant. Recommender systems give personalized has shown that explanation has been found more valuable by the recommendation to the user by either using explicit data provided user if they are explained in a more simple and accurate by user through ratings or by using implicit data like user manner[1]. User motivation data is inferred from previous user browsing behavior, past purchasing behavior etc. The popularity check-in and comments at different locations. We also show how of personalized systems have increased manifold as today the our approach can be applied through a case study on a real life success of e-commerce sites is dependent on the quality of dataset of 10 users extracted from popular location based recommendations. Hence, researchers are continuously trying to recommendation app Foursquare. improve quality of recommendation by integrating more and more data about the customers in the recommendation process [1]. 2. RECOMMENDER SYSTEMS IN Presently, there is a clear trend towards usage of context-aware TOURISM recommendation systems as they integrate contextual data like The key problems in recommender systems are the prediction time, location, mood, emotions, companion, purpose etc. with problem and the top-N prediction problem[5]. The prediction ratings data to provide final recommendation[2]. Among the problem is about predicting whether a user will like or dislike a different contexts, research community has shown most interest new item that the user has not yet consumed or purchased. This towards location-aware recommendations systems. One reason for prediction is generated using the knowledge of user preferences, greater focus on location-aware recommendation systems is the past purchases data and interests. The top-N problem in easy availability of GPS data due to increased adoption of smart recommender systems attempts to predict the set of N items that a user may like from the set of items he has not yet seen. Recommender systems in tourism industry primarily focuses on Copyright held by the author(s). the top-N problem. In tourism industry these systems help the RecTour 2016 - Workshop on Recommenders in Tourism held in tourist or user in information search by recommending conjunction with the 10th ACM Conference on Recommender Systems (RecSys), September 15, 2016, Boston, MA, USA. destinations, point of interests, restaurants, events, travel Pure check-in data approach: This approach primarily considers itineraries etc. The recommendations made are specific to a user check-in frequency data for making recommendations. It assumes as they are personalized according to the user interests and that if two users are similar if they have similar checked in preferences. history. One demerit in considering check in data frequency as ratings is that during vacations tourists only check in once at a The popularity of recommender systems in tourism industry has tourist location so it difficult to deduce whether the user liked or brought this field into the attention of the academic research disliked the place. community. The increased focus on research in recommender systems in tourism is evident by going through the detailed and Geographical influenced approach: The current location of the exhaustive survey papers [6], [7] that have been published on the user and distance of POIs not yet visited by the user from the topic recently. Among the recommendation problems that are current location is used for making recommendations. This researched in the tourism domain, Point of Interests approach is appropriate when availability of time, transport recommendations (POI) is the most researched problem by the options, traffic condition, weather conditions are used as academic community [6]. contextual variables for making recommendation. In Point of Interests recommendations a ranked list of point of Social influence enhanced approach: Popularity of location based interests like tourists attractions in a city, restaurants, events etc. social networks like Foursquare, Yelp etc. have resulted in are presented to the user[8]–[11]. POI problem can be classified recommendation approaches that utilize social relationships as top-N recommendation problem. These systems focus on two among users to enhance POI recommendation. This approach aspects of the problem, first on how to improve accuracy of the assumes that friends of a user have similar interests as the user recommendations and the other aspect is how to effectively and a user is more likely expected to trust recommendations made present the information to the user[8]. Majority of recommender by people who they are connected to in the network. systems in tourism focus on point of interests recommendations. One primary reason for that is the availability of new contextual Temporal influence enhanced approach: Some POIs are preferred data that has motivated researchers to focus on ways to improve to be visited at a particular time slot, temporal influence approach recommendation accuracy. Location, time of the day, current considers time information while generating recommendations. weather, budget, means of transport, traffic, presence of friends For example, there are tourist locations that are primarily visited nearby etc. [6]are contextual aspects that have been used in during sunrise or sunset time. Even closing time and opening time making POI recommendations. Location of the user is one the of museums and restaurants are important information that can most popular contextual data that is used in most algorithms, one help improve POIs recommendation. reason could be the easy access to accurate location data because Sequential influenced approach: These systems assume that users of widespread use of mobile phones among tourists. Social exhibits pattern in the order in which they visit places. For network data is also used for making POI recommendations[10]– example, some users may prefer going to a restaurant after [12]. Social network data provides rich data points that can be watching a movie or a game in a stadium. Patterns once identified used for profiling the user. It also provides data about relationship from past check in data can be used for making recommendations. between users, preferences and views that can be derived from user comments, reviews and other network activity. Categorical influenced approach: Users preferences for checking in at particular categories of point of interests is leveraged in this Tour Package [13] or Travel destination recommendation and approach. A user may prefer going to museums only and another Itinerary Planning [14], [15] are two more problems that have user may have preferences for entertainment parks. The been researched. Travel destination recommendations are knowledge of a user biases for a particular category of POI is used designed with tour operators as end users. These systems also in this approach for enhancing recommendations. recommend hotels, flights in addition to tourist locations. Cost is also one aspect that is considered an important criteria in tour Among the different approaches for POIs recommendations, recommendations[13]. Itinerary planning or route planning check-in data, geographical influenced and temporal influenced recommends multiple day personalized tour plans with set of approach have significantly enhanced POI recommendation point of interests to be explored each day. Contextual aspects like quality. Geographical influence is used the most to improve POI days of visit, pace of travel, preferred transportation mode [16] recommendation [17]. have been used for such recommendations. In [11] a approach is proposed that combines temporal and Among the recommender systems approaches in tourism domain geographical data to make POI recommendation. Their approach research, content based technique is the more popular as splits time into hourly slots and mines the user checking in history compared to collaborative filtering technique [6]. Unavailability to get insight about user temporal preferences to visit particular of user rating data for different attractions, restaurants, events etc. type of POIs at a time slot. As users tend to visit POIs that are may be the reason behind fewer collaborative filtering based closer to their current location, this approach combines the POIs approaches. Hybrid algorithms that combine content based and nearby to user location with the insight acquired by the user collaborative filtering based may be considered more appropriate temporal preferences to make the final recommendation. for tourism domain recommendations. Social network data, geographical data as well as check in data is used in the approach proposed in [19]. Their approach challenge 3. RELATED WORK the main assumption made in most POIs recommendations Point of interest recommendations approaches in context based approaches that use location based social network (LBSN) data recommender systems is categorized by the type of data the i.e. check-in frequency of user at a particular POI indicates user systems process to make recommendations [17], [18]. Combining preference for that POI. This assumption is challenged on the both the categorization approaches, POI recommendation basis that in more than 50 % of the places a user has checked in approaches can be of six types. only once and on the basis of one check in it cannot be implied that the user prefers that POI. In the approach proposed by [19], Information Motivation is commonly observed in youth, usually a they extract the preference of POI by mining user comments for suggestion or advice. Social Motivation is used when hanging out that POI. The mining of the comments provide a sentiment with friends or for relationship development. polarity for the POI for that user. The sentiment polarity can be Entertainment value is when user is relaxing or playing, to positive, negative or neutral. The final recommendation is made communicate positive moments. by integrating user sentiment polarity towards POIs he has commented on, user social network links and geographical Gameful experience is using gaming mechanics in non-gaming location of the user. sense. City spots and achieving a virtual status like Mayor or owner. Utilitarian motivation is for winning promotions and Most approaches use geographical data, check-in data, and discounts as you share or check-in at a place. temporal data or combine them to make recommendations. An interesting approach [20] uses user personality data to enhance the Belongingness is for places like home, school when users are recommendations. The personality of the user is captured through nostalgic. a questionnaire filled by the user during the registration process Scenario: Number of places a tourist can visit is limited because on the mobile application. The personality is based on the Five of the constraints of time and effort needed. POIs recommender factor model [21]. The Five Factor Model terms personality systems help the users in deciding the POIs to visit using among the five dimensions of Extraversion, Agreeableness, contextual variables. The final list to 2-3 POIs provided to the Conscientiousness, Neuroticism, and Openness to Experience. user as recommendation many times are difficult to justify as Along with personality of the user the approach uses a set of multiple contextual variables are evaluated using complex contextual factors, such as the weather conditions, the time of day, algorithms to generate the final recommendations. In our user’s location and user’s mood to recommend the final set of approach we further refine the final recommendations based on POIs. user motivation to checking in at a POI. The justification of the Our approach uses the concept of user motivation for checking in recommendations made through explanations based on user as the context to refine the final recommendation. To the best of motivation for checking in will be easier for the user to our knowledge, no other research paper has ever used this data for comprehend. POIs recommendations. For example, a tourist in Barcelona whose analysis of checking in data in Foursquare suggests that he is motivated by social 4. MOTIVATION BASED enhancement will be recommended POIs like Sagrada Familia or RECOMMENDATION APPROACH Park Guell, while somebody who is motivated by information 4.1 Motivation motivation will be recommended an offbeat attraction or a new Spatiotemporal mobility among user using location based social restaurant. networks (LBSN) are driven primarily by social rewards and also by systems rewards [22]. Checking in behavior in LBSN is driven 4.2 Algorithm Our aim is to recommend User Ui at location Li a place of interest by users seeking status recognition in his network .LSBN enables Pi that is within a radius of distance Ri from location Li. We define social recognition as the feature of immediate sharing of location two kind of motivations for each location or POI and for each details, pictures etc. generates immediate social reaction among user. The two motivations are Dominant explicit motivation and his network friends. Checking in behavior is an important aspect Dominant perceived motivation. Dominant explicit motivation for in marketing of services in LBSN. The authors cite the theory of a user is derived from explicit data like comments and status self-concept [23] to explain the behavior of customers. Theory of messages after checking in at a POI on the location based social self-concept indicates that consumers value consumption that network. Dominant perceived motivation are generated for a results in recognition and that strengthen the conception about location through survey. themselves. Similarly, we use motivation behind checking in at a We use the approach of explicit and perceived motivation because location to refine recommendations as we believe that every user many users may not put any comments or status messages after may have a different motivation behind checking in at a location. checking in at a location. Using explicit motivation will more Using user motivation preferences while showing and explaining likely result in data sparsity. the final POIs recommendation to the users will result in more effective recommendations. Step 1: Assigning dominant explicit motivations to users and Our work is based on the foundation that users have a particular locations motivation when they check-in at a location. In this work, we use Dominant explicit motivations for a user are determined based on the classification done by [24], they found that motivations for a the motivation inferred from the comments and status messages user to share his location or check-in at a particular location can user have given after check in to different places. Set DUi be classified into seven categories. represents the dominant motivations of a user Ui .It contains those They identify Social Enhancement, Informational Motive, Social motivations which have highest frequency of check-ins with a Motivation, Entertainment value, Gameful Experience, Utilitarian particular motivation. We have made DUi a set as a user may motivation, Belongingness as the motives for a user to check-in at have more than one motivation having the max frequency count. a location. Similarly, Dominant explicit motivations to a place is referred as set DPi and is determined by doing a frequency count of the Social enhancement value is the most commonly observed inferred motivations derived from comments given to the place by motive, exhibited in more than fifty percent check-ins, where a users. user check-ins for impressing others and feels important to be at a place [25]. Step 2: Assigning dominant perceived motivations to users and locations. Based on offline assessment of the places by a survey each place the foursquare were selected for the final analysis that has more Pi is assigned a perceived motivation. PPi is the set of dominant than 10 check-in in Indore. We could find 10 users with such perceived motivations of a place Pi. It is determined by doing a criteria who had visited in all 97 places including restaurants, frequency count of the perceived motivations assigned to the pubs, city spots, home and business. place Pi in the survey. PUi is the set of dominant perceived motivations for a user Ui. It is determined by doing a frequency 5.2 Comment Classification count of the perceived motivations assigned to each place the user The 7 motivations for check-in by [24] are used, Table 1 shows Ui has checked into. which characteristics of a comment can help us map with which motivation. For example, if a user checks-in at a high end Step 3: Recommendation Generation restaurant and puts a comment “Tremendous food”. Then his To recommend User Ui at location Li a place of interest Pi that is motivation would be classified as social enhancement value as it within a radius of distance Ri from location Li.. Using is a high end restaurant and the user has checked in as he is collaborative filtering or other POIs recommendation algorithm feeling important. Based on his comment the user motivation will approaches a set of places within a radius of distance Ri from be classified as information motivation. Similarly, all the location Li. are generated that are matching with user preferences comments by the user are classified by using characteristics of the based on his ratings or preferences data. motivation. Table 2 shows the result of classifying all the 129 comments made by the users in our dataset. The table shows the Step 4: Final set of motivation based recommendation distribution of various motivations. User Ui set of dominant motivations as generated in step 2 is the union of the sets DUi and PUi. Place Pi set of dominant 5.3 User Classification motivations as generated in step 2 is the union of the sets DPi and Every user has one motivation from the above 7 categories. The PPi. Then the final set of recommendations is based on refining motivation of the user is the highest frequency of motivation in the places selected in step 3 using User Ui dominant motivations. the comments as classified according to the above method. Hence, From the set of places selected in step 3 only those places Pi a user Ui has a motivation Mi, if the comments posted by the user whose dominant motivations matches with user Ui dominant on foursquare has highest number of comments with Mi as motivations are recommended to the user Ui. motivation. In our dataset of 10 four square users in Indore, 50 per cent had Social Enhancement value as their main motivation. Our proposed algorithm approach applies post filtering contextual What was surprising was that both social enhancement and approach [2] as motivation context is applied on a list of Informational motivation together were dominant motivation in recommendations generated by traditional recommender systems 20 per cent user. Hence, for a user it is not necessary to have a algorithms. A pre-filtering contextual approach can also be single motivation as a dominant motivation but combination of applied but as ratings data is primarily used by traditional more than one. Table 3 shows classification of users on the algorithms, pre-filtering places of interest based on motivations basis of 7 motivations. may lead to data sparsity problem. Table 1. Characteristics of different motivations for location 5. Case Study check-ins Our approach as mentioned in the earlier section is to refine the Motivation Characteristics recommendation made by an algorithm that is designed for accuracy. Our suggested approach objective is not to improve Social Enhancement Impressing others accuracy further but to improve the way final recommendations Feeling Important are explained to the user. Explanations[26] are an important To show off component of recommender systems as it may increase the Extremely Popular location adaptability and trustworthiness of the recommender system. In Night clubs [27], the authors show that there is merit in providing High end restaurants personalized explanations and explanation interfaces should be Distinctive Identity or Intellectual designed to increase the informativeness of the explanation. We Image believe our approach will add to the informativeness of the Celebrity Status explanation. Informational Motivation Suggestions Advices Instead of an experimental evaluation of our approach we have Information about event or news done a data analysis on four square data set to check whether our Location and arrival approach is feasible in a real life scenario. Our approach is Important event feasible only if users show variety of motivation while checking Give and take recommendation in, if all users show the same motivation then motivation cannot be used to refine the final recommendations. Our algorithm uses Social Motivation Meeting new people the concept of perceived and actual motivation, we also want to Socializing check through actual data whether there is any difference in actual Observing others and perceived motivation. Meeting a Friend Flirting and relationships 5.1 Data Collection Emotional Feeling Foursquare launched in 2009 is used for check-in and real time At Home or Office location sharing with friends. It has 50 million users in its network Know about friends and where and handles millions of check-in in a day. The Foursquare app they are allows the users to have their own profile and share their comment Entertainment Value Playing describing their feelings when they visit a location. The users of Relaxing Passing Time 5.4 Perceived & Actual Motivation Less Lonely While the user giving a comment on a location he visits might be Positive moments, emotional state classified into one of the motivation category, but this motivation Fight boredom may differ for the perceived dominant motivation of the location. Initiate chat The perceived dominant motivation of the location is classified Waiting based on a survey. This mismatch in perceived and actual Gameful Experience To collect award Points motivation in check-in can lead to distorted image of the user. For Status in an app example, suppose a user checks-in at a high end posh restaurant More in females with a comment “Excellent coffee, Must try.” Though, the actual Older age( not in youth aged 19- motivation of the user is Information Motivation but the 22) characteristics of the place may make another user who sees this City spots(streets, square, roads, comment assume the motivation behind check-in was Social bridge, old town) Enhancement Value. To address this dissonance, in step 4 of the Escape from reality, Virtual algorithm, for a User Ui, the set of dominant motivations is Possession generated by the union of the sets DUi and PUi. We analyzed the Utilitarian Win promotions and discounts data to check whether this kind of dissonance exists in our data Check in of family business for set. Table 4 shows 39% of times the actual motivation is also the marketing perceived motivation but a majority number of times the Belongingness Place with social group perceived and actual motivation differs. Also, 12 percent places Nostalgia or ownership had multiple classifications which dint allow us to attach them to a specific motivation. Table 2. Comments of Four Square dataset classified on Table 4. Difference between actual and perceived motivation motivation for check-in Perceived & Places Percentage Actual Motivation Per cent of comments Motivation Social Enhancement Value 38 Equal 35 39.32584 Informational Motivation 27 Not Equal 43 48.31461 Social Motivation 7.7 Not Determined 11 12.35955 Entertainment Value 10.0 Gameful experiences 3.8 6. DISCUSSION & CONCLUSION Every user has a motivation when the user checks-in at a particular location, if these motivation is taken into account while Utilitarian Motivation 3.8 generating final recommendations then it will be more beneficial to the user. Context variables like time, location and social Belongingness 9.30 network data of a user are mainly used to recommend new locations to a tourist. In this paper, we propose an approach that uses user checking in motivation along with the other contextual variables. Motivation can be effectively used if used as a post- Table 3. Spread of Motivation among users in our dataset filtering contextual variable in combination with the existing Motivation Per-cent recommendation algorithm. Our analysis of real life data shows that our approach can be used as user do show different Social Enhancement Value 50 motivations as they check in into different POIs , the primary motivation among users also differs and there do exist a difference between a user’s actual motivation for checking in and Informational Motivation 10 perceived motivation for checking in. We believe using our approach will improve the explanation quality of the final Social Enhancement Value and recommendations. Informational 20 Limitations of the study are that we did not experimentally evaluate the accuracy of our approach based on metrics like mean Motivation Social Motivation 10 absolute error, precision or recall. Our approach is not designed to improve accuracy, what it offers is the additional explanation for a Belongingness 10 recommendation to the user, which helps him understand the recommendation given more easily. In future research we aim to operationalize this algorithm on a mobile app and then try to do a qualitative evaluation of the ability of the algorithm in providing more satisfaction to the user. Though, the existing dataset is [14] Y. Kurata and T. Hara, ‘CT-Planner4: Toward a More sufficient for gaining insights on the appropriateness of our User-Friendly Interactive Day-Tour Planner’, in algorithm, a qualitative study is required to show the benefit of Information and Communication Technologies in Tourism using checking in motivation for enhancing POIs 2014, Z. Xiang and I. Tussyadiah, Eds. Cham: Springer recommendations. International Publishing, 2013, pp. 73–86. [15] R. Schaller and D. Elsweiler, ‘Itinerary recommenders: 7. REFERENCES how do users customize their routes and what can we learn [1] J. Bobadilla, F. Ortega, A. Hernando, and A. Gutiérrez, from them?’, 2014, pp. 185–194. ‘Recommender systems survey’, Knowl.-Based Syst., vol. [16] A. Garcia, P. Vansteenwegen, O. Arbelaitz, W. Souffriau, 46, pp. 109–132, Jul. 2013. and M. T. Linaza, ‘Integrating public transportation in [2] G. Adomavicius and A. Tuzhilin, ‘Context-Aware personalised electronic tourist guides’, Comput. Oper. Res., Recommender Systems’, in Recommender Systems vol. 40, no. 3, pp. 758–774, Mar. 2013. Handbook, F. Ricci, L. Rokach, B. Shapira, and P. B. [17] Y. Yu and X. Chen, ‘A survey of point-of-interest Kantor, Eds. Boston, MA: Springer US, 2011, pp. 217– recommendation in location-based social networks’, 253. presented at the Workshops at the Twenty-Ninth AAAI [3] J. E. Dickinson, K. Ghali, T. Cherrett, C. Speed, N. Davies, Conference on Artificial Intelligence, 2015. and S. Norgate, ‘Tourism and the smartphone app: [18] J.-D. Zhang and C.-Y. Chow, ‘Point-of-interest capabilities, emerging practice and scope in the travel recommendations in location-based social networks’, domain’, Curr. Issues Tour., vol. 17, no. 1, pp. 84–101, SIGSPATIAL Spec., vol. 7, no. 3, pp. 26–33, Jan. 2016. Jan. 2014. [19] J.-D. Zhang, C.-Y. Chow, and Y. Zheng, ‘ORec: An [4] I. P. Tussyadiah, ‘The Influence of Innovativeness on On- Opinion-Based Point-of-Interest Recommendation Site Smartphone Use Among American Travelers: Framework’, presented at the Proceedings of the 24th Implications for Context-Based Push Marketing’, J. Travel ACM International on Conference on Information and Tour. Mark., vol. 33, no. 6, pp. 806–823, Jul. 2016. Knowledge Management, 2015, pp. 1641–1650. [5] M. Deshpande and G. Karypis, ‘Item-based top- N [20] M. Braunhofer, M. Elahi, and F. Ricci, ‘Usability recommendation algorithms’, ACM Trans. Inf. Syst., vol. Assessment of a Context-Aware and Personality-Based 22, no. 1, pp. 143–177, Jan. 2004. Mobile Recommender System’, presented at the [6] J. Borràs, A. Moreno, and A. Valls, ‘Intelligent tourism International Conference on Electronic Commerce and recommender systems: A survey’, Expert Syst. Appl., vol. Web Technologies, 2014, pp. 77–88. 41, no. 16, pp. 7370–7389, Nov. 2014. [21] R. R. McCrae and O. P. John, ‘An Introduction to the Five- [7] D. Gavalas, C. Konstantopoulos, K. Mastakas, and G. Factor Model and Its Applications’, J. Pers., vol. 60, no. 2, Pantziou, ‘Mobile recommender systems in tourism’, J. pp. 175–215, Jun. 1992. Netw. Comput. Appl., vol. 39, pp. 319–333, Mar. 2014. [22] I. P. Tussyadiah, ‘A Concept of Location-Based Social [8] J. M. Noguera, M. J. Barranco, R. J. Segura, and L. Network Marketing’, J. Travel Tour. Mark., vol. 29, no. 3, Martínez, ‘A mobile 3D-GIS hybrid recommender system pp. 205–220, Apr. 2012. for tourism’, Inf. Sci., vol. 215, pp. 37–52, Dec. 2012. [23] E. L. Grubb and H. L. Grathwohl, ‘Consumer Self- [9] F. Ricci, ‘Travel recommender systems’, IEEE Intell. Syst., Concept, Symbolism and Market Behavior: A Theoretical 2002. Approach’, J. Mark., vol. 31, no. 4, pp. 22–27, 1967. [10] J. J.-C. Ying, E. H.-C. Lu, W.-N. Kuo, and V. S. Tseng, [24] R. Yavuz and A. Toker, ‘Location sharing on social ‘Urban point-of-interest recommendation by mining user networks: implications for marketing’, Mark. Intell. Plan., check-in behaviors’, 2012, p. 63. vol. 32, no. 5, pp. 567–585, Aug. 2014. [11] Q. Yuan, G. Cong, Z. Ma, A. Sun, and N. M.- Thalmann, [25] U. M. Dholakia, R. P. Bagozzi, and L. K. Pearo, ‘A social ‘Time-aware point-of-interest recommendation’, 2013, p. influence model of consumer participation in network- and 363. small-group-based virtual communities’, Int. J. Res. Mark., [12] N. Saiph Savage, M. Baranski, N. Elva Chavez, and T. vol. 21, no. 3, pp. 241–263, Sep. 2004. Höllerer, ‘I’m feeling LoCo: A Location Based Context [26] N. Tintarev, ‘Explanations of recommendations’, 2007, p. Aware Recommendation System’, in Advances in 203. Location-Based Services, G. Gartner and F. Ortag, Eds. [27] A. Sharma and D. Cosley, ‘Do social explanations work?: Berlin, Heidelberg: Springer Berlin Heidelberg, 2012, pp. studying and modeling the effects of social explanations in 37–54. recommender systems’, 2013, pp. 1133–1144. [13] Y. Ge, H. Xiong, A. Tuzhilin, and Q. Liu, ‘Cost-Aware Collaborative Filtering for Travel Tour Recommendations’, ACM Trans. Inf. Syst., vol. 32, no. 1, pp. 1–31, Jan. 2014.